api: Fix params_show for untracked targets.

If an untracked target is provided alongside `stages` or `deps`, don't load the parameters from the target and raise no params found instead.

Closes #8161

.git_archival.txt

@@ -1 +1,4 @@
+node: $Format:%H$
+node-date: $Format:%cI$
+describe-name: $Format:%(describe:tags=true)$
 ref-names: $Format:%D$

.github/FUNDING.yml

@@ -1,3 +0,0 @@
-# These are supported funding model platforms
-
-patreon: DVCorg

.github/workflows/benchmarks.yaml

@@ -11,12 +11,12 @@ jobs:
         os: [ubuntu-18.04]
     steps:
       - name: Cancel Previous Runs
-        uses: styfle/cancel-workflow-action@0.9.1
+        uses: styfle/cancel-workflow-action@0.10.0
         with:
           access_token: ${{ github.token }}
-      - uses: actions/setup-python@v3
+      - uses: actions/setup-python@v4
         with:
-            python-version: 3.7
+            python-version: 3.8
       - uses: iterative/dvc-bench@main
         with:
             pytest_options: "-k 'test_init or test_help'"

.github/workflows/packages.yaml

@@ -6,152 +6,92 @@ on:
   release:
     types: [released, prereleased]
   workflow_dispatch:
+
 jobs:
-  deb:
-    runs-on: ubuntu-18.04
+  binary:
+    strategy:
+      matrix:
+        include:
+         - {pkg: "deb",  os: "ubuntu-18.04", asset: "dvc_${{ github.event.release.tag_name }}_amd64.deb"}
+         - {pkg: "rpm", os: "ubuntu-18.04", asset: "dvc-${{ github.event.release.tag_name }}-1.x86_64.rpm"}
+         - {pkg: "osxpkg", os: "macos-10.15", asset: "dvc-${{ github.event.release.tag_name }}.pkg"}
+         - {pkg: "exe",  os: "windows-2019", asset: "dvc-${{ github.event.release.tag_name }}.exe"}
+
+    name: ${{ matrix.pkg }}
+    runs-on: ${{ matrix.os }}
     steps:
     - uses: actions/checkout@v3
       with:
           fetch-depth: 0
+
     - name: Set up Python 3.9
-      uses: actions/setup-python@v3
+      uses: actions/setup-python@v4
       with:
         python-version: 3.9
-    - name: Set up Ruby 2.4
-      uses: actions/[email protected]
+        cache: pip
+        cache-dependency-path: |
+          setup.cfg
+          scripts/build-requirements.txt
+
+    - name: Set up Ruby 2.6
+      uses: ruby/setup-ruby@v1
+      if: matrix.pkg != 'exe'
       with:
-        ruby-version: 2.4
-    - name: Install deps
-      run: |
-        pip install .[all]
-        pip install -r scripts/build-requirements.txt
-        gem install --no-document fpm
-    - name: Build deb
-      run: python scripts/build.py deb
-    - name: Publish deb
-      if: github.event_name == 'release'
-      uses: actions/[email protected]
-      env:
-        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      with:
-        upload_url: ${{ github.event.release.upload_url }}
-        asset_path: scripts/fpm/dvc_${{ github.event.release.tag_name }}_amd64.deb
-        asset_name: dvc_${{ github.event.release.tag_name }}_amd64.deb
-        asset_content_type: binary/octet-stream
-  rpm:
-    runs-on: ubuntu-18.04
-    steps:
-    - uses: actions/checkout@v3
-      with:
-        fetch-depth: 0
-    - name: Set up Python 3.9
-      uses: actions/setup-python@v3
-      with:
-        python-version: 3.9
-    - name: Set up Ruby 2.4
-      uses: actions/[email protected]
-      with:
-        ruby-version: 2.4
-    - name: Install deps
-      run: |
-        pip install .[all]
-        pip install -r scripts/build-requirements.txt
-        gem install --no-document fpm
-    - name: Build rpm
-      run: python scripts/build.py rpm
-    - name: Publish rpm
-      if: github.event_name == 'release'
-      uses: actions/[email protected]
-      env:
-        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      with:
-        upload_url: ${{ github.event.release.upload_url }}
-        asset_path: scripts/fpm/dvc-${{ github.event.release.tag_name }}-1.x86_64.rpm
-        asset_name: dvc-${{ github.event.release.tag_name }}-1.x86_64.rpm
-        asset_content_type: binary/octet-stream
-  osxpkg:
-    runs-on: macos-10.15
-    steps:
-    - uses: actions/checkout@v3
-      with:
-        fetch-depth: 0
-    - name: Set up Python 3.9
-      uses: actions/setup-python@v3
-      with:
-        python-version: 3.9
-    - name: Set up Ruby 2.4
-      uses: actions/[email protected]
-      with:
-        ruby-version: 2.4
-    - name: Install deps
-      run: |
-        # available lxml (webdav dependency) wheels are built with an old
-        # SDK, which breaks osxpkg notarization. Building from scratch
-        # instead.
-        pip install .[all] --no-binary lxml
-        pip install -r scripts/build-requirements.txt
-        gem install --no-document fpm
-    - name: Build
-      run: python scripts/build.py osxpkg
-    - name: Publish
-      if: github.event_name == 'release'
-      uses: actions/[email protected]
-      env:
-        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      with:
-        upload_url: ${{ github.event.release.upload_url }}
-        asset_path: scripts/fpm/dvc-${{ github.event.release.tag_name }}.pkg
-        asset_name: dvc-${{ github.event.release.tag_name }}.pkg
-        asset_content_type: binary/octet-stream
-  exe:
-    runs-on: windows-2019
-    steps:
-    - uses: actions/checkout@v3
-      with:
-        fetch-depth: 0
-    - name: Set up Python 3.9
-      uses: actions/setup-python@v3
-      with:
-        python-version: 3.9
+        ruby-version: '2.6'
+
+    - name: Install fpm
+      if: matrix.pkg != 'exe'
+      run: gem install --no-document fpm
+
     - name: Install deps
       run: |
+        pip install -U wheel pip setuptools
         pip install .[all]
         pip install -r scripts/build-requirements.txt
+
     - name: Pull images
-      run: |
-        dvc pull
-    - name: Build
-      run: python scripts\build.py exe
-      shell: cmd
-    - name: Publish
+      run: dvc pull
+
+    - name: Build ${{ matrix.pkg }}
+      run: python scripts/build.py ${{ matrix.pkg }}
+
+    - name: Publish ${{ matrix.pkg }}
       if: github.event_name == 'release'
+      uses: actions/[email protected]
       env:
         GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-      uses: actions/[email protected]
       with:
         upload_url: ${{ github.event.release.upload_url }}
-        asset_path: scripts/innosetup/dvc-${{ github.event.release.tag_name }}.exe
-        asset_name: dvc-${{ github.event.release.tag_name }}.exe
+        asset_path: scripts/${{ matrix.pkg == 'exe' && 'innosetup' || 'fpm' }}/${{ matrix.asset }}
+        asset_name: ${{ matrix.asset }}
         asset_content_type: binary/octet-stream
+
   pip:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
     steps:
     - uses: actions/checkout@v3
       with:
         fetch-depth: 0
-    - name: Set up Python 3.7
-      uses: actions/setup-python@v3
+
+    - name: Set up Python 3.8
+      uses: actions/setup-python@v4
       with:
-        python-version: 3.7
+        python-version: 3.8
+        cache: pip
+        cache-dependency-path: setup.cfg
+
     - name: Install
-      run: pip install ".[all,tests]"
+      run: |
+        pip install -U wheel pip build setuptools
+        pip install ".[all,tests]"
+
     - name: Build packages
       run: |
         pip install -U pip
         ./scripts/build_package.sh
+
     - name: Publish packages to PyPI
       if: github.event_name == 'release'
-      uses: pypa/gh-action-pypi-publish@master
+      uses: pypa/gh-action-pypi-publish@release/v1
       with:
-        user: __token__
         password: ${{ secrets.PYPI_TOKEN }}

.github/workflows/tests.yaml

@@ -17,14 +17,14 @@ jobs:
     runs-on: ubuntu-latest
     steps:
     - name: Cancel Previous Runs
-      uses: styfle/cancel-workflow-action@0.9.1
+      uses: styfle/cancel-workflow-action@0.10.0
       with:
         access_token: ${{ github.token }}
     - uses: actions/checkout@v3
       with:
         fetch-depth: 0
     - name: Set up Python 3.8
-      uses: actions/setup-python@v3
+      uses: actions/setup-python@v4
       with:
         python-version: 3.8
     - name: Install requirements
@@ -34,7 +34,7 @@ jobs:
         pip install ".[dev]" pre-commit
     - name: Check README
       run: python setup.py checkdocs
-    - uses: pre-commit/action@v2.0.3
+    - uses: pre-commit/action@v3.0.0
   tests:
     timeout-minutes: 45
     runs-on: ${{ matrix.os }}
@@ -42,13 +42,13 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-20.04, windows-latest, macos-latest]
-        pyv: ["3.7", "3.8", "3.9", "3.10"]
+        pyv: ["3.8", "3.9", "3.10"]
     steps:
     - uses: actions/checkout@v3
       with:
         fetch-depth: 0
     - name: Set up Python
-      uses: actions/setup-python@v3
+      uses: actions/setup-python@v4
       with:
         python-version: ${{ matrix.pyv }}
     - name: get pip cache dir

.pre-commit-config.yaml

@@ -6,9 +6,9 @@ repos:
       - id: black
         language_version: python3
     repo: https://github.com/ambv/black
-    rev: 22.3.0
+    rev: 22.6.0
   - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v4.2.0
+    rev: v4.3.0
     hooks:
       - id: check-toml
       - id: check-yaml

README.rst

@@ -3,97 +3,114 @@
 `Website <https://dvc.org>`_
 • `Docs <https://dvc.org/doc>`_
 • `Blog <http://blog.dataversioncontrol.com>`_
-• `Twitter <https://twitter.com/DVCorg>`_
-• `Chat (Community & Support) <https://dvc.org/chat>`_
 • `Tutorial <https://dvc.org/doc/get-started>`_
-• `Mailing List <https://sweedom.us10.list-manage.com/subscribe/post?u=a08bf93caae4063c4e6a351f6&id=24c0ecc49a>`_
+• `Related Technologies <https://dvc.org/doc/user-guide/related-technologies>`_
+• `How DVC works`_
+• `VS Code Extension`_
+• `Installation`_
+• `Contributing`_
+• `Community and Support`_
 
-|CI| |Maintainability| |Coverage| |Donate| |DOI|
+|CI| |Python Version| |Coverage| |VS Code| |DOI|
 
 |PyPI| |Packages| |Brew| |Conda| |Choco| |Snap|
 
 |
 
-**Data Version Control** or **DVC** is an **open-source** tool for data science and machine
-learning projects. Key features:
+**Data Version Control** or **DVC** is a command line tool and `VS Code Extension`_ to help you develop reproducible machine learning projects:
 
-#. Simple **command line** Git-like experience. Does not require installing and maintaining
-   any databases. Does not depend on any proprietary online services.
+#. **Version** your data and models.
+   Store them in your cloud storage but keep their version info in your Git repo.
 
-#. Management and versioning of **datasets** and **machine learning
-   models**. Data can be saved in S3, Google cloud, Azure, Alibaba 
-   cloud, SSH server, HDFS, or even local HDD RAID.
+#. **Iterate** fast with lightweight pipelines.
+   When you make changes, only run the steps impacted by those changes.
 
-#. Makes projects **reproducible** and **shareable**; helping to answer questions about how
-   a model was built.
+#. **Track** experiments in your local Git repo (no servers needed).
 
-#. Helps manage experiments with Git tags/branches and **metrics** tracking.
+#. **Compare** any data, code, parameters, model, or performance plots.
 
-**DVC** aims to replace spreadsheet and document sharing tools (such as Excel or Google Docs)
-frequently used as both knowledge repositories and team ledgers.
-DVC also replaces both ad-hoc scripts to track, move, and deploy different model versions
-and ad-hoc data file suffixes and prefixes.
-
-.. contents:: **Contents**
-  :backlinks: none
-
-How DVC works
-=============
-
-We encourage you to read our `Get Started <https://dvc.org/doc/get-started>`_ guide to better understand what DVC
-is and how it can fit your scenarios.
-
-The easiest (but not perfect!) *analogy* to describe it: DVC is Git (or Git-LFS to be precise) & Makefiles
-made right and tailored specifically for ML and Data Science scenarios.
-
-#. ``Git/Git-LFS`` part - DVC helps store and share data artifacts and models, connecting them with a Git repository.
-#. ``Makefile``\ s part - DVC describes how one data or model artifact was built from other data and code.
-
-DVC usually runs along with Git. Git is used as usual to store and version code (including DVC meta-files). DVC helps
-to store data and model files seamlessly out of Git, while preserving almost the same user experience as if they
-were stored in Git itself. To store and share the data cache, DVC supports multiple remotes - any cloud (S3, Azure,
-Google Cloud, etc) or any on-premise network storage (via SSH, for example).
-
-|Flowchart|
-
-The DVC pipelines (computational graph) feature connects code and data together. It is possible to explicitly
-specify all steps required to produce a model: input dependencies including data, commands to run,
-and output information to be saved. See the quick start section below or
-the `Get Started <https://dvc.org/doc/get-started>`_ tutorial to learn more.
+#. **Share** experiments and automatically reproduce anyone's experiment.
 
 Quick start
 ===========
 
-Please read `Get Started <https://dvc.org/doc/get-started>`_ guide for a full version. Common workflow commands include:
+    Please read our `Command Reference <https://dvc.org/doc/command-reference>`_ for a complete list.
 
-+-----------------------------------+----------------------------------------------------------------------------+
-| Step                              | Command                                                                    |
-+===================================+============================================================================+
-| Track data                        | | ``$ git add train.py``                                                   |
-|                                   | | ``$ dvc add images.zip``                                                 |
-+-----------------------------------+----------------------------------------------------------------------------+
-| Connect code and data by commands | | ``$ dvc run -n prepare -d images.zip -o images/ unzip -q images.zip``    |
-|                                   | | ``$ dvc run -n train -d images/ -d train.py -o model.p python train.py`` |
-+-----------------------------------+----------------------------------------------------------------------------+
-| Make changes and reproduce        | | ``$ vi train.py``                                                        |
-|                                   | | ``$ dvc repro model.p.dvc``                                              |
-+-----------------------------------+----------------------------------------------------------------------------+
-| Share code                        | | ``$ git add .``                                                          |
-|                                   | | ``$ git commit -m 'The baseline model'``                                 |
-|                                   | | ``$ git push``                                                           |
-+-----------------------------------+----------------------------------------------------------------------------+
-| Share data and ML models          | | ``$ dvc remote add myremote -d s3://mybucket/image_cnn``                 |
-|                                   | | ``$ dvc push``                                                           |
-+-----------------------------------+----------------------------------------------------------------------------+
+A common CLI workflow includes:
+
+
++-----------------------------------+----------------------------------------------------------------------------------------------------+
+| Task                              | Terminal                                                                                           |
++===================================+====================================================================================================+
+| Track data                        | | ``$ git add train.py params.yaml``                                                               |
+|                                   | | ``$ dvc add images/``                                                                            |
++-----------------------------------+----------------------------------------------------------------------------------------------------+
+| Connect code and data             | | ``$ dvc stage add -n featurize -d images/ -o features/ python featurize.py``                     |
+|                                   | | ``$ dvc stage add -n train -d features/ -d train.py -o model.p -M metrics.json python train.py`` |
++-----------------------------------+----------------------------------------------------------------------------------------------------+
+| Make changes and experiment       | | ``$ dvc exp run -n exp-baseline``                                                                |
+|                                   | | ``$ vi train.py``                                                                                |
+|                                   | | ``$ dvc exp run -n exp-code-change``                                                             |
++-----------------------------------+----------------------------------------------------------------------------------------------------+
+| Compare and select experiments    | | ``$ dvc exp show``                                                                               |
+|                                   | | ``$ dvc exp apply exp-baseline``                                                                 |
++-----------------------------------+----------------------------------------------------------------------------------------------------+
+| Share code                        | | ``$ git add .``                                                                                  |
+|                                   | | ``$ git commit -m 'The baseline model'``                                                         |
+|                                   | | ``$ git push``                                                                                   |
++-----------------------------------+----------------------------------------------------------------------------------------------------+
+| Share data and ML models          | | ``$ dvc remote add myremote -d s3://mybucket/image_cnn``                                         |
+|                                   | | ``$ dvc push``                                                                                   |
++-----------------------------------+----------------------------------------------------------------------------------------------------+
+
+How DVC works
+=============
+
+    We encourage you to read our `Get Started
+    <https://dvc.org/doc/get-started>`_ docs to better understand what DVC
+    does and how it can fit your scenarios.
+
+The closest *analogies* to describe the main DVC features are these:
+
+#. **Git for data**: Store and share data artifacts (like Git-LFS but without a server) and models, connecting them with a Git repository. Data management meets GitOps!
+#. **Makefiles** for ML: Describes how data or model artifacts are built from other data and code in a standard format. Now you can version your data pipelines with Git.
+#. Local **experiment tracking**: Turn your machine into an ML experiment management platform, and collaborate with others using existing Git hosting (Github, Gitlab, etc.).
+
+Git is employed as usual to store and version code (including DVC meta-files as placeholders for data).
+DVC `stores data and model files <https://dvc.org/doc/start/data-management>`_ seamlessly in a cache outside of Git, while preserving almost the same user experience as if they were in the repo.
+To share and back up the *data cache*, DVC supports multiple remote storage platforms - any cloud (S3, Azure, Google Cloud, etc.) or on-premise network storage (via SSH, for example).
+
+|Flowchart|
+
+`DVC pipelines <https://dvc.org/doc/start/data-management/pipelines>`_ (computational graphs) connect code and data together.
+They specify all steps required to produce a model: input dependencies including code, data, commands to run; and output information to be saved.
+
+Last but not least, `DVC Experiment Versioning <https://dvc.org/doc/start/experiments>`_ lets you prepare and run a large number of experiments.
+Their results can be filtered and compared based on hyperparameters and metrics, and visualized with multiple plots.
+
+.. _`VS Code Extension`:
+
+Visual Studio Code Extension
+============================
+
+|VS Code|
+
+To use DVC as a GUI right from your VS Code IDE, install the `DVC Extension <https://marketplace.visualstudio.com/items?itemName=Iterative.dvc>`_ from the Marketplace.
+It currently features experiment tracking and data management, and more features (data pipeline support, etc.) are coming soon!
+
+|VS Code Extension Overview|
+
+    Note: You'll have to install core DVC on your system separately (as detailed
+    below). The Extension will guide you if needed.
 
 Installation
 ============
 
-There are four options to install DVC: ``pip``, Homebrew, Conda (Anaconda) or an OS-specific package.
+There are several ways to install DVC: in VS Code; using ``snap``, ``choco``, ``brew``, ``conda``, ``pip``; or with an OS-specific package.
 Full instructions are `available here <https://dvc.org/doc/get-started/install>`_.
 
-Snap (Snapcraft/Linux)
-----------------------
+Snapcraft (Linux)
+-----------------
 
 |Snap|
 
@@ -102,11 +119,10 @@ Snap (Snapcraft/Linux)
    snap install dvc --classic
 
 This corresponds to the latest tagged release.
-Add ``--beta`` for the latest tagged release candidate,
-or ``--edge`` for the latest ``main`` version.
+Add ``--beta`` for the latest tagged release candidate, or ``--edge`` for the latest ``main`` version.
 
-Choco (Chocolatey/Windows)
---------------------------
+Chocolatey (Windows)
+--------------------
 
 |Choco|
 
@@ -114,8 +130,8 @@ Choco (Chocolatey/Windows)
 
    choco install dvc
 
-Brew (Homebrew/Mac OS)
-----------------------
+Brew (mac OS)
+-------------
 
 |Brew|
 
@@ -123,8 +139,8 @@ Brew (Homebrew/Mac OS)
 
    brew install dvc
 
-Conda (Anaconda)
-----------------
+Anaconda (Any platform)
+-----------------------
 
 |Conda|
 
@@ -133,11 +149,10 @@ Conda (Anaconda)
    conda install -c conda-forge mamba # installs much faster than conda
    mamba install -c conda-forge dvc
 
-Depending on the remote storage type you plan to use to keep and share your data, you might need to
-install optional dependencies: `dvc-s3`, `dvc-azure`, `dvc-gdrive`, `dvc-gs`, `dvc-oss`, `dvc-ssh`.
+Depending on the remote storage type you plan to use to keep and share your data, you might need to install optional dependencies: `dvc-s3`, `dvc-azure`, `dvc-gdrive`, `dvc-gs`, `dvc-oss`, `dvc-ssh`.
 
-pip (PyPI)
-----------
+PyPI (Python)
+-------------
 
 |PyPI|
 
@@ -145,10 +160,8 @@ pip (PyPI)
 
    pip install dvc
 
-Depending on the remote storage type you plan to use to keep and share your data, you might need to specify
-one of the optional dependencies: ``s3``, ``gs``, ``azure``, ``oss``, ``ssh``. Or ``all`` to include them all.
-The command should look like this: ``pip install dvc[s3]`` (in this case AWS S3 dependencies such as ``boto3``
-will be installed automatically).
+Depending on the remote storage type you plan to use to keep and share your data, you might need to specify one of the optional dependencies: ``s3``, ``gs``, ``azure``, ``oss``, ``ssh``. Or ``all`` to include them all.
+The command should look like this: ``pip install 'dvc[s3]'`` (in this case AWS S3 dependencies such as ``boto3`` will be installed automatically).
 
 To install the development version, run:
 
@@ -156,88 +169,58 @@ To install the development version, run:
 
    pip install git+git://github.com/iterative/dvc
 
-Package
--------
+Package (Platform-specific)
+---------------------------
 
 |Packages|
 
-Self-contained packages for Linux, Windows, and Mac are available. The latest version of the packages
-can be found on the GitHub `releases page <https://github.com/iterative/dvc/releases>`_.
+Self-contained packages for Linux, Windows, and Mac are available.
+The latest version of the packages can be found on the GitHub `releases page <https://github.com/iterative/dvc/releases>`_.
 
 Ubuntu / Debian (deb)
 ^^^^^^^^^^^^^^^^^^^^^
 .. code-block:: bash
 
    sudo wget https://dvc.org/deb/dvc.list -O /etc/apt/sources.list.d/dvc.list
-   sudo apt-get update
-   sudo apt-get install dvc
+   wget -qO - https://dvc.org/deb/iterative.asc | sudo apt-key add -
+   sudo apt update
+   sudo apt install dvc
 
 Fedora / CentOS (rpm)
 ^^^^^^^^^^^^^^^^^^^^^
 .. code-block:: bash
 
    sudo wget https://dvc.org/rpm/dvc.repo -O /etc/yum.repos.d/dvc.repo
+   sudo rpm --import https://dvc.org/rpm/iterative.asc
    sudo yum update
    sudo yum install dvc
 
-Comparison to related technologies
-==================================
-
-#. Data Engineering tools such as `AirFlow <https://airflow.apache.org/>`_,
-   `Luigi <https://github.com/spotify/luigi>`_, and others - in DVC data,
-   model and ML pipelines represent a single ML project focused on data
-   scientists' experience.  Data engineering tools orchestrate multiple data
-   projects and focus on efficient execution. A DVC project can be used from
-   existing data pipelines as a single execution step.
-
-#. `Git-annex <https://git-annex.branchable.com/>`_ - DVC uses the idea of storing the content of large files (which should
-   not be in a Git repository) in a local key-value store, and uses file hardlinks/symlinks instead of
-   copying/duplicating files.
-
-#. `Git-LFS <https://git-lfs.github.com/>`_ - DVC is compatible with many
-   remote storage services (S3, Google Cloud, Azure, SSH, etc). DVC also
-   uses reflinks or hardlinks to avoid copy operations on checkouts; thus
-   handling large data files much more efficiently.
-
-#. Makefile (and analogues including ad-hoc scripts) - DVC tracks
-   dependencies (in a directed acyclic graph).
-
-#. `Workflow Management Systems <https://en.wikipedia.org/wiki/Workflow_management_system>`_ - DVC is a workflow
-   management system designed specifically to manage machine learning experiments. DVC is built on top of Git.
-
-#. `DAGsHub <https://dagshub.com/>`_ - online service to host DVC
-   projects.  It provides a useful UI around DVC repositories and integrates
-   other tools.
-
-#. `DVC Studio <https://studio.iterative.ai/>`_ - official online
-   platform for DVC projects.  It can be used to manage data and models, run
-   and track experiments, and visualize and share results.  Also, it
-   integrates with `CML (CI/CD for ML) <https://cml.dev/>`__ for training
-   models in the cloud or Kubernetes.
-
-
 Contributing
 ============
 
-|Maintainability| |Donate|
+|Maintainability|
 
-Contributions are welcome! Please see our `Contributing Guide <https://dvc.org/doc/user-guide/contributing/core>`_ for more
-details. Thanks to all our contributors!
+Contributions are welcome!
+Please see our `Contributing Guide <https://dvc.org/doc/user-guide/contributing/core>`_ for more details.
+Thanks to all our contributors!
 
 |Contribs|
 
-Mailing List
-============
+Community and Support
+=====================
 
-Want to stay up to date? Want to help improve DVC by participating in our occasional polls? Subscribe to our `mailing list <https://sweedom.us10.list-manage.com/subscribe/post?u=a08bf93caae4063c4e6a351f6&id=24c0ecc49a>`_. No spam, really low traffic.
+* `Twitter <https://twitter.com/DVCorg>`_
+* `Forum <https://discuss.dvc.org/>`_
+* `Discord Chat <https://dvc.org/chat>`_
+* `Email <mailto:[email protected]>`_
+* `Mailing List <https://sweedom.us10.list-manage.com/subscribe/post?u=a08bf93caae4063c4e6a351f6&id=24c0ecc49a>`_
 
 Copyright
 =========
 
 This project is distributed under the Apache license version 2.0 (see the LICENSE file in the project root).
 
-By submitting a pull request to this project, you agree to license your contribution under the Apache license version
-2.0 to this project.
+By submitting a pull request to this project, you agree to license your contribution under the Apache license version 2.0 to this project.
 
 Citation
 ========
@@ -254,6 +237,9 @@ Barrak, A., Eghan, E.E. and Adams, B. `On the Co-evolution of ML Pipelines and S
    :target: https://dvc.org
    :alt: DVC logo
 
+.. |VS Code Extension Overview| image:: https://raw.githubusercontent.com/iterative/vscode-dvc/main/extension/docs/overview.gif
+   :alt: DVC Extension for VS Code
+
 .. |CI| image:: https://github.com/iterative/dvc/workflows/Tests/badge.svg?branch=main
    :target: https://github.com/iterative/dvc/actions
    :alt: GHA Tests
@@ -262,14 +248,14 @@ Barrak, A., Eghan, E.E. and Adams, B. `On the Co-evolution of ML Pipelines and S
    :target: https://codeclimate.com/github/iterative/dvc
    :alt: Code Climate
 
+.. |Python Version| image:: https://img.shields.io/pypi/pyversions/dvc
+   :target: https://pypi.org/project/dvc
+   :alt: Python Version
+
 .. |Coverage| image:: https://codecov.io/gh/iterative/dvc/branch/main/graph/badge.svg
    :target: https://codecov.io/gh/iterative/dvc
    :alt: Codecov
 
-.. |Donate| image:: https://img.shields.io/badge/patreon-donate-green.svg?logo=patreon
-   :target: https://www.patreon.com/DVCorg/overview
-   :alt: Donate
-
 .. |Snap| image:: https://img.shields.io/badge/snap-install-82BEA0.svg?logo=snapcraft
    :target: https://snapcraft.io/dvc
    :alt: Snapcraft
@@ -305,3 +291,7 @@ Barrak, A., Eghan, E.E. and Adams, B. `On the Co-evolution of ML Pipelines and S
 .. |Contribs| image:: https://contrib.rocks/image?repo=iterative/dvc
    :target: https://github.com/iterative/dvc/graphs/contributors
    :alt: Contributors
+
+.. |VS Code| image:: https://vsmarketplacebadge.apphb.com/version/Iterative.dvc.svg
+   :target: https://marketplace.visualstudio.com/items?itemName=Iterative.dvc
+   :alt: VS Code Extension

dvc/analytics.py

@@ -50,7 +50,7 @@ def is_enabled():
             Config(validate=False).get("core", {}).get("analytics", "true")
         )
 
-    logger.debug("Analytics is {}abled.".format("en" if enabled else "dis"))
+    logger.debug("Analytics is %sabled.", "en" if enabled else "dis")
 
     return enabled
 
@@ -187,4 +187,4 @@ def _find_or_create_user_id():
             return user_id
 
     except LockError:
-        logger.debug(f"Failed to acquire '{lockfile}'")
+        logger.debug("Failed to acquire '%s'", lockfile)

dvc/api.py

@@ -1,123 +0,0 @@
-import os
-from contextlib import _GeneratorContextManager as GCM
-
-from funcy import reraise
-
-from dvc.exceptions import OutputNotFoundError, PathMissingError
-from dvc.repo import Repo
-
-
-def get_url(path, repo=None, rev=None, remote=None):
-    """
-    Returns the URL to the storage location of a data file or directory tracked
-    in a DVC repo. For Git repos, HEAD is used unless a rev argument is
-    supplied. The default remote is tried unless a remote argument is supplied.
-
-    Raises OutputNotFoundError if the file is not tracked by DVC.
-
-    NOTE: This function does not check for the actual existence of the file or
-    directory in the remote storage.
-    """
-    with Repo.open(repo, rev=rev, subrepos=True, uninitialized=True) as _repo:
-        with reraise(FileNotFoundError, PathMissingError(path, repo)):
-            info = _repo.repo_fs.info(path)
-
-        dvc_info = info.get("dvc_info")
-        if not dvc_info:
-            raise OutputNotFoundError(path, repo)
-
-        dvc_repo = info["repo"]
-        md5 = dvc_info["md5"]
-
-        return dvc_repo.cloud.get_url_for(remote, checksum=md5)
-
-
-def open(  # noqa, pylint: disable=redefined-builtin
-    path, repo=None, rev=None, remote=None, mode="r", encoding=None
-):
-    """
-    Open file in the supplied path tracked in a repo (both DVC projects and
-    plain Git repos are supported). For Git repos, HEAD is used unless a rev
-    argument is supplied. The default remote is tried unless a remote argument
-    is supplied. It may only be used as a context manager:
-
-        with dvc.api.open(
-                'path/to/file',
-                repo='https://example.com/url/to/repo'
-                ) as fd:
-            # ... Handle file object fd
-    """
-    args = (path,)
-    kwargs = {
-        "repo": repo,
-        "remote": remote,
-        "rev": rev,
-        "mode": mode,
-        "encoding": encoding,
-    }
-    return _OpenContextManager(_open, args, kwargs)
-
-
-class _OpenContextManager(GCM):
-    def __init__(
-        self, func, args, kwds
-    ):  # pylint: disable=super-init-not-called
-        self.gen = func(*args, **kwds)
-        self.func, self.args, self.kwds = func, args, kwds
-
-    def __getattr__(self, name):
-        raise AttributeError(
-            "dvc.api.open() should be used in a with statement."
-        )
-
-
-def _open(path, repo=None, rev=None, remote=None, mode="r", encoding=None):
-    with Repo.open(repo, rev=rev, subrepos=True, uninitialized=True) as _repo:
-        with _repo.open_by_relpath(
-            path, remote=remote, mode=mode, encoding=encoding
-        ) as fd:
-            yield fd
-
-
-def read(path, repo=None, rev=None, remote=None, mode="r", encoding=None):
-    """
-    Returns the contents of a tracked file (by DVC or Git). For Git repos, HEAD
-    is used unless a rev argument is supplied. The default remote is tried
-    unless a remote argument is supplied.
-    """
-    with open(
-        path, repo=repo, rev=rev, remote=remote, mode=mode, encoding=encoding
-    ) as fd:
-        return fd.read()
-
-
-def make_checkpoint():
-    """
-    Signal DVC to create a checkpoint experiment.
-
-    If the current process is being run from DVC, this function will block
-    until DVC has finished creating the checkpoint. Otherwise, this function
-    will return immediately.
-    """
-    import builtins
-    from time import sleep
-
-    from dvc.env import DVC_CHECKPOINT, DVC_ROOT
-    from dvc.stage.monitor import CheckpointTask
-
-    if os.getenv(DVC_CHECKPOINT) is None:
-        return
-
-    root_dir = os.getenv(DVC_ROOT, Repo.find_root())
-    signal_file = os.path.join(
-        root_dir, Repo.DVC_DIR, "tmp", CheckpointTask.SIGNAL_FILE
-    )
-
-    with builtins.open(signal_file, "w", encoding="utf-8") as fobj:
-        # NOTE: force flushing/writing empty file to disk, otherwise when
-        # run in certain contexts (pytest) file may not actually be written
-        fobj.write("")
-        fobj.flush()
-        os.fsync(fobj.fileno())
-    while os.path.exists(signal_file):
-        sleep(0.1)

dvc/api/__init__.py

@@ -0,0 +1,9 @@
+from .data import (  # noqa, pylint: disable=redefined-builtin
+    get_url,
+    open,
+    read,
+)
+from .experiments import make_checkpoint
+from .params import params_show
+
+__all__ = ["get_url", "make_checkpoint", "open", "params_show", "read"]

dvc/api/data.py

@@ -0,0 +1,213 @@
+from contextlib import _GeneratorContextManager as GCM
+from typing import Optional
+
+from funcy import reraise
+
+from dvc.exceptions import OutputNotFoundError, PathMissingError
+from dvc.repo import Repo
+
+
+def get_url(path, repo=None, rev=None, remote=None):
+    """
+    Returns the URL to the storage location of a data file or directory tracked
+    in a DVC repo. For Git repos, HEAD is used unless a rev argument is
+    supplied. The default remote is tried unless a remote argument is supplied.
+
+    Raises OutputNotFoundError if the file is not tracked by DVC.
+
+    NOTE: This function does not check for the actual existence of the file or
+    directory in the remote storage.
+    """
+    with Repo.open(repo, rev=rev, subrepos=True, uninitialized=True) as _repo:
+        fs_path = _repo.dvcfs.from_os_path(path)
+        with reraise(FileNotFoundError, PathMissingError(path, repo)):
+            info = _repo.dvcfs.info(fs_path)
+
+        dvc_info = info.get("dvc_info")
+        if not dvc_info:
+            raise OutputNotFoundError(path, repo)
+
+        dvc_repo = info["repo"]
+        md5 = dvc_info["md5"]
+
+        return dvc_repo.cloud.get_url_for(remote, checksum=md5)
+
+
+class _OpenContextManager(GCM):
+    def __init__(
+        self, func, args, kwds
+    ):  # pylint: disable=super-init-not-called
+        self.gen = func(*args, **kwds)
+        self.func, self.args, self.kwds = func, args, kwds
+
+    def __getattr__(self, name):
+        raise AttributeError(
+            "dvc.api.open() should be used in a with statement."
+        )
+
+
+def open(  # noqa, pylint: disable=redefined-builtin
+    path: str,
+    repo: Optional[str] = None,
+    rev: Optional[str] = None,
+    remote: Optional[str] = None,
+    mode: str = "r",
+    encoding: Optional[str] = None,
+):
+    """
+    Opens a file tracked in a DVC project.
+
+    This function may only be used as a context manager (using the `with`
+    keyword, as shown in the examples).
+
+    This function makes a direct connection to the remote storage, so the file
+    contents can be streamed. Your code can process the data buffer as it's
+    streamed, which optimizes memory usage.
+
+    Note:
+        Use dvc.api.read() to load the complete file contents
+        in a single function call, no context manager involved.
+        Neither function utilizes disc space.
+
+    Args:
+        path (str): location and file name of the target to open,
+        relative to the root of `repo`.
+        repo (str, optional): location of the DVC project or Git Repo.
+            Defaults to the current DVC project (found by walking up from the
+            current working directory tree).
+            It can be a URL or a file system path.
+            Both HTTP and SSH protocols are supported for online Git repos
+            (e.g. [user@]server:project.git).
+        rev (str, optional): Any `Git revision`_ such as a branch or tag name,
+            a commit hash or a dvc experiment name.
+            Defaults to HEAD.
+            If `repo` is not a Git repo, this option is ignored.
+        remote (str, optional): Name of the `DVC remote`_ used to form the
+            returned URL string.
+            Defaults to the `default remote`_ of `repo`.
+            For local projects, the cache is tried before the default remote.
+        mode (str, optional): Specifies the mode in which the file is opened.
+            Defaults to "r" (read).
+            Mirrors the namesake parameter in builtin `open()`_.
+            Only reading `mode` is supported.
+        encoding(str, optional): `Codec`_ used to decode the file contents.
+            Defaults to None.
+            This should only be used in text mode.
+            Mirrors the namesake parameter in builtin `open()`_.
+
+    Returns:
+        _OpenContextManager: A context manager that generatse a corresponding
+            `file object`_.
+            The exact type of file object depends on the mode used.
+            For more details, please refer to Python's `open()`_ built-in,
+            which is used under the hood.
+
+    Raises:
+        AttributeError: If this method is not used as a context manager.
+        ValueError: If non-read `mode` is used.
+
+    Examples:
+
+        - Use data or models from a DVC repository.
+
+        Any file tracked in a DVC project (and stored remotely) can be
+        processed directly in your Python code with this API.
+        For example, an XML file tracked in a public DVC repo on GitHub can be
+        processed like this:
+
+        >>> from xml.sax import parse
+        >>> import dvc.api
+        >>> from mymodule import mySAXHandler
+
+        >>> with dvc.api.open(
+        ...     'get-started/data.xml',
+        ...     repo='https://github.com/iterative/dataset-registry'
+        ... ) as fd:
+        ...     parse(fd, mySAXHandler)
+
+        We use a SAX XML parser here because dvc.api.open() is able to stream
+        the data from remote storage.
+        The mySAXHandler object should handle the event-driven parsing of the
+        document in this case.
+        This increases the performance of the code (minimizing memory usage),
+        and is typically faster than loading the whole data into memory.
+
+        - Accessing private repos
+
+        This is just a matter of using the right repo argument, for example an
+        SSH URL (requires that the credentials are configured locally):
+
+        >>> import dvc.api
+
+        >>> with dvc.api.open(
+        ...     'features.dat',
+        ...     repo='[email protected]:path/to/repo.git'
+        ... ) as fd:
+        ...     # ... Process 'features'
+        ...     pass
+
+        - Use different versions of data
+
+        Any git revision (see `rev`) can be accessed programmatically.
+        For example, if your DVC repo has tagged releases of a CSV dataset:
+
+        >>> import csv
+        >>> import dvc.api
+        >>> with dvc.api.open(
+        ...     'clean.csv',
+        ...     rev='v1.1.0'
+        ... ) as fd:
+        ...     reader = csv.reader(fd)
+        ...     # ... Process 'clean' data from version 1.1.0
+
+    .. _Git revision:
+        https://git-scm.com/docs/revisions
+
+    .. _DVC remote:
+        https://dvc.org/doc/command-reference/remote
+
+    .. _default remote:
+        https://dvc.org/doc/command-reference/remote/default
+
+    .. _open():
+        https://docs.python.org/3/library/functions.html#open
+
+    .. _Codec:
+        https://docs.python.org/3/library/codecs.html#standard-encodings
+
+    .. _file object:
+        https://docs.python.org/3/glossary.html#term-file-object
+
+    """
+    if "r" not in mode:
+        raise ValueError("Only reading `mode` is supported.")
+
+    args = (path,)
+    kwargs = {
+        "repo": repo,
+        "remote": remote,
+        "rev": rev,
+        "mode": mode,
+        "encoding": encoding,
+    }
+    return _OpenContextManager(_open, args, kwargs)
+
+
+def _open(path, repo=None, rev=None, remote=None, mode="r", encoding=None):
+    with Repo.open(repo, rev=rev, subrepos=True, uninitialized=True) as _repo:
+        with _repo.open_by_relpath(
+            path, remote=remote, mode=mode, encoding=encoding
+        ) as fd:
+            yield fd
+
+
+def read(path, repo=None, rev=None, remote=None, mode="r", encoding=None):
+    """
+    Returns the contents of a tracked file (by DVC or Git). For Git repos, HEAD
+    is used unless a rev argument is supplied. The default remote is tried
+    unless a remote argument is supplied.
+    """
+    with open(
+        path, repo=repo, rev=rev, remote=remote, mode=mode, encoding=encoding
+    ) as fd:
+        return fd.read()

dvc/api/experiments.py

@@ -0,0 +1,33 @@
+import builtins
+import os
+from time import sleep
+
+from dvc.env import DVC_CHECKPOINT, DVC_ROOT
+from dvc.repo import Repo
+from dvc.stage.monitor import CheckpointTask
+
+
+def make_checkpoint():
+    """
+    Signal DVC to create a checkpoint experiment.
+
+    If the current process is being run from DVC, this function will block
+    until DVC has finished creating the checkpoint. Otherwise, this function
+    will return immediately.
+    """
+    if os.getenv(DVC_CHECKPOINT) is None:
+        return
+
+    root_dir = os.getenv(DVC_ROOT, Repo.find_root())
+    signal_file = os.path.join(
+        root_dir, Repo.DVC_DIR, "tmp", CheckpointTask.SIGNAL_FILE
+    )
+
+    with builtins.open(signal_file, "w", encoding="utf-8") as fobj:
+        # NOTE: force flushing/writing empty file to disk, otherwise when
+        # run in certain contexts (pytest) file may not actually be written
+        fobj.write("")
+        fobj.flush()
+        os.fsync(fobj.fileno())
+    while os.path.exists(signal_file):
+        sleep(0.1)

dvc/api/params.py

@@ -0,0 +1,274 @@
+from collections import Counter
+from typing import Dict, Iterable, Optional, Union
+
+from funcy import first
+
+from dvc.exceptions import DvcException
+from dvc.repo import Repo
+
+
+def params_show(
+    *targets: str,
+    repo: Optional[str] = None,
+    stages: Optional[Union[str, Iterable[str]]] = None,
+    rev: Optional[str] = None,
+    deps: bool = False,
+) -> Dict:
+    """Get parameters tracked in `repo`.
+
+    Without arguments, this function will retrieve all params from all tracked
+    parameter files, for the current working tree.
+
+    See the options below to restrict the parameters retrieved.
+
+    Args:
+        *targets (str, optional): Names of the parameter files to retrieve
+        params from. For example, "params.py, myparams.toml".
+        If no `targets` are provided, all parameter files tracked in `dvc.yaml`
+        will be used.
+        Note that targets don't necessarily have to be defined in `dvc.yaml`.
+        repo (str, optional): location of the DVC repository.
+            Defaults to the current project (found by walking up from the
+            current working directory tree).
+            It can be a URL or a file system path.
+            Both HTTP and SSH protocols are supported for online Git repos
+            (e.g. [user@]server:project.git).
+        stages (Union[str, Iterable[str]], optional): Name or names of the
+            stages to retrieve parameters from.
+            Defaults to `None`.
+            If `None`, all parameters from all stages will be retrieved.
+        rev (str, optional): Name of the `Git revision`_ to retrieve parameters
+            from.
+            Defaults to `None`.
+            An example of git revision can be a branch or tag name, a commit
+            hash or a dvc experiment name.
+            If `repo` is not a Git repo, this option is ignored.
+            If `None`, the current working tree will be used.
+        deps (bool, optional): Whether to retrieve only parameters that are
+            stage dependencies or not.
+            Defaults to `False`.
+
+    Returns:
+        Dict: See Examples below.
+
+    Raises:
+        DvcException: If no params are found in `repo`.
+
+    Examples:
+
+        - No arguments.
+
+        Working on https://github.com/iterative/example-get-started
+
+        >>> import json
+        >>> import dvc.api
+        >>> params = dvc.api.params_show()
+        >>> print(json.dumps(params, indent=4))
+        {
+            "prepare": {
+                "split": 0.2,
+                "seed": 20170428
+            },
+            "featurize": {
+                "max_features": 200,
+                "ngrams": 2
+            },
+            "train": {
+                "seed": 20170428,
+                "n_est": 50,
+                "min_split": 0.01
+            }
+        }
+
+        ---
+
+        - Filtering with `stages`.
+
+        Working on https://github.com/iterative/example-get-started
+
+        `stages` can a single string:
+
+        >>> import json
+        >>> import dvc.api
+        >>> params = dvc.api.params_show(stages="prepare")
+        >>> print(json.dumps(params, indent=4))
+        {
+            "prepare": {
+                "split": 0.2,
+                "seed": 20170428
+            }
+        }
+
+        Or an iterable of strings:
+
+        >>> import json
+        >>> import dvc.api
+        >>> params = dvc.api.params_show(stages=["prepare", "train"])
+        >>> print(json.dumps(params, indent=4))
+        {
+            "prepare": {
+                "split": 0.2,
+                "seed": 20170428
+            },
+            "train": {
+                "seed": 20170428,
+                "n_est": 50,
+                "min_split": 0.01
+            }
+        }
+
+        ---
+
+        - Using `rev`.
+
+        Working on https://github.com/iterative/example-get-started
+
+        >>> import json
+        >>> import dvc.api
+        >>> params = dvc.api.params_show(rev="tune-hyperparams")
+        >>> print(json.dumps(params, indent=4))
+        {
+            "prepare": {
+                "split": 0.2,
+                "seed": 20170428
+            },
+            "featurize": {
+                "max_features": 200,
+                "ngrams": 2
+            },
+            "train": {
+                "seed": 20170428,
+                "n_est": 100,
+                "min_split": 8
+            }
+        }
+
+        ---
+
+        - Using `targets`.
+
+        Working on `multi-params-files` folder of
+        https://github.com/iterative/pipeline-conifguration
+
+        You can pass a single target:
+
+        >>> import json
+        >>> import dvc.api
+        >>> params = dvc.api.params_show("params.yaml")
+        >>> print(json.dumps(params, indent=4))
+        {
+            "run_mode": "prod",
+            "configs": {
+                "dev": "configs/params_dev.yaml",
+                "test": "configs/params_test.yaml",
+                "prod": "configs/params_prod.yaml"
+            },
+            "evaluate": {
+                "dataset": "micro",
+                "size": 5000,
+                "metrics": ["f1", "roc-auc"],
+                "metrics_file": "reports/metrics.json",
+                "plots_cm": "reports/plot_confusion_matrix.png"
+            }
+        }
+
+
+        Or multiple targets:
+
+        >>> import json
+        >>> import dvc.api
+        >>> params = dvc.api.params_show(
+        ...     "configs/params_dev.yaml", "configs/params_prod.yaml")
+        >>> print(json.dumps(params, indent=4))
+        {
+            "configs/params_prod.yaml:run_mode": "prod",
+            "configs/params_prod.yaml:config_file": "configs/params_prod.yaml",
+            "configs/params_prod.yaml:data_load": {
+                "dataset": "large",
+                "sampling": {
+                "enable": true,
+                "size": 50000
+                }
+            },
+            "configs/params_prod.yaml:train": {
+                "epochs": 1000
+            },
+            "configs/params_dev.yaml:run_mode": "dev",
+            "configs/params_dev.yaml:config_file": "configs/params_dev.yaml",
+            "configs/params_dev.yaml:data_load": {
+                "dataset": "development",
+                "sampling": {
+                "enable": true,
+                "size": 1000
+                }
+            },
+            "configs/params_dev.yaml:train": {
+                "epochs": 10
+            }
+        }
+
+        ---
+
+        - Git URL as `repo`.
+
+        >>> import json
+        >>> import dvc.api
+        >>> params = dvc.api.params_show(
+        ...     repo="https://github.com/iterative/demo-fashion-mnist")
+        {
+            "train": {
+                "batch_size": 128,
+                "hidden_units": 64,
+                "dropout": 0.4,
+                "num_epochs": 10,
+                "lr": 0.001,
+                "conv_activation": "relu"
+            }
+        }
+
+
+    .. _Git revision:
+        https://git-scm.com/docs/revisions
+
+    """
+    if isinstance(stages, str):
+        stages = [stages]
+
+    def _onerror_raise(result: Dict, exception: Exception, *args, **kwargs):
+        raise exception
+
+    def _postprocess(params):
+        processed = {}
+        for rev, rev_data in params.items():
+            processed[rev] = {}
+
+            counts = Counter()
+            for file_data in rev_data["data"].values():
+                for k in file_data["data"]:
+                    counts[k] += 1
+
+            for file_name, file_data in rev_data["data"].items():
+                to_merge = {
+                    (k if counts[k] == 1 else f"{file_name}:{k}"): v
+                    for k, v in file_data["data"].items()
+                }
+                processed[rev] = {**processed[rev], **to_merge}
+
+        if "workspace" in processed:
+            del processed["workspace"]
+
+        if not processed:
+            raise DvcException("No params found")
+
+        return processed[first(processed)]
+
+    with Repo.open(repo) as _repo:
+        params = _repo.params.show(
+            revs=rev if rev is None else [rev],
+            targets=targets,
+            deps=deps,
+            onerror=_onerror_raise,
+            stages=stages,
+        )
+
+    return _postprocess(params)

dvc/cli/__init__.py

@@ -6,6 +6,7 @@ import sys
 # Workaround for CPython bug. See [1] and [2] for more info.
 # [1] https://github.com/aws/aws-cli/blob/1.16.277/awscli/clidriver.py#L55
 # [2] https://bugs.python.org/issue29288
+from typing import Optional
 
 "".encode("idna")
 
@@ -36,6 +37,100 @@ def parse_args(argv=None):
     return args
 
 
+def _log_unknown_exceptions() -> None:
+    from dvc.info import get_dvc_info
+    from dvc.logger import FOOTER
+
+    logger.exception("unexpected error")
+    if logger.isEnabledFor(logging.DEBUG):
+        logger.debug("Version info for developers:\n%s", get_dvc_info())
+    logger.info(FOOTER)
+    return None
+
+
+def _log_exceptions(exc: Exception) -> Optional[int]:
+    """Try to log some known exceptions, that are not DVCExceptions."""
+    from dvc.utils import error_link, format_link
+
+    if isinstance(exc, OSError):
+        import errno
+
+        if exc.errno == errno.EMFILE:
+            logger.exception(
+                "too many open files, please visit "
+                "{} to see how to handle this "
+                "problem".format(error_link("many-files")),
+                extra={"tb_only": True},
+            )
+        else:
+            _log_unknown_exceptions()
+        return None
+
+    from dvc.fs import AuthError, ConfigError, RemoteMissingDepsError
+
+    if isinstance(exc, RemoteMissingDepsError):
+        from dvc.utils.pkg import PKG
+
+        proto = exc.protocol
+        by_pkg = {
+            "pip": f"pip install 'dvc[{proto}]'",
+            "conda": f"conda install -c conda-forge dvc-{proto}",
+        }
+
+        cmd = by_pkg.get(PKG)
+        if cmd:
+            link = format_link("https://dvc.org/doc/install")
+            hint = (
+                f"To install dvc with those dependencies, run:\n"
+                "\n"
+                f"\t{cmd}\n"
+                "\n"
+                f"See {link} for more info."
+            )
+        else:
+            link = format_link("https://github.com/iterative/dvc/issues")
+            hint = f"\nPlease report this bug to {link}. Thank you!"
+
+        logger.exception(
+            f"URL '{exc.url}' is supported but requires these missing "
+            f"dependencies: {exc.missing_deps}. {hint}",
+            extra={"tb_only": True},
+        )
+        return None
+
+    if isinstance(exc, (AuthError, ConfigError)):
+        link = format_link("https://man.dvc.org/remote/modify")
+        logger.exception("configuration error")
+        logger.exception(
+            f"{exc!s}\nLearn more about configuration settings at {link}.",
+            extra={"tb_only": True},
+        )
+        return 251
+
+    from dvc_data.hashfile.cache import DiskError
+
+    if isinstance(exc, DiskError):
+        from dvc.utils import relpath
+
+        directory = relpath(exc.directory)
+        logger.exception(
+            f"Could not open pickled '{exc.type}' cache.\n"
+            f"Remove the '{directory}' directory and then retry this command."
+            f"\nSee {error_link('pickle')} for more information.",
+            extra={"tb_only": True},
+        )
+        return None
+
+    from dvc_data.build import IgnoreInCollectedDirError
+
+    if isinstance(exc, IgnoreInCollectedDirError):
+        logger.exception("")
+        return None
+
+    _log_unknown_exceptions()
+    return None
+
+
 def main(argv=None):  # noqa: C901
     """Main entry point for dvc CLI.
 
@@ -48,7 +143,7 @@ def main(argv=None):  # noqa: C901
     from dvc._debug import debugtools
     from dvc.config import ConfigError
     from dvc.exceptions import DvcException, NotDvcRepoError
-    from dvc.logger import FOOTER, disable_other_loggers
+    from dvc.logger import disable_other_loggers, set_loggers_level
 
     # NOTE: stderr/stdout may be closed if we are running from dvc.daemon.
     # On Linux we directly call cli.main after double forking and closing
@@ -76,7 +171,7 @@ def main(argv=None):  # noqa: C901
             level = logging.TRACE
 
         if level is not None:
-            logger.setLevel(level)
+            set_loggers_level(level)
 
         logger.trace(args)
 
@@ -94,6 +189,15 @@ def main(argv=None):  # noqa: C901
     except KeyboardInterrupt:
         logger.exception("interrupted by the user")
         ret = 252
+    except BrokenPipeError:
+        import os
+
+        # Python flushes standard streams on exit; redirect remaining output
+        # to devnull to avoid another BrokenPipeError at shutdown
+        # See: https://docs.python.org/3/library/signal.html#note-on-sigpipe
+        devnull = os.open(os.devnull, os.O_WRONLY)
+        os.dup2(devnull, sys.stdout.fileno())
+        ret = 141  # 128 + 13 (SIGPIPE)
     except NotDvcRepoError:
         logger.exception("")
         ret = 253
@@ -104,27 +208,7 @@ def main(argv=None):  # noqa: C901
         ret = 254
     except Exception as exc:  # noqa, pylint: disable=broad-except
         # pylint: disable=no-member
-        import errno
-
-        if isinstance(exc, OSError) and exc.errno == errno.EMFILE:
-            from dvc.utils import error_link
-
-            logger.exception(
-                "too many open files, please visit "
-                "{} to see how to handle this "
-                "problem".format(error_link("many-files")),
-                extra={"tb_only": True},
-            )
-        else:
-            from dvc.info import get_dvc_info
-
-            logger.exception("unexpected error")
-
-            dvc_info = get_dvc_info()
-            logger.debug("Version info for developers:\n%s", dvc_info)
-
-            logger.info(FOOTER)
-        ret = 255
+        ret = _log_exceptions(exc) or 255
 
     try:
         from dvc import analytics

dvc/cli/command.py

@@ -13,7 +13,7 @@ class CmdBase(ABC):
 
         os.chdir(args.cd)
 
-        self.repo = Repo(uninitialized=self.UNINITIALIZED)
+        self.repo: "Repo" = Repo(uninitialized=self.UNINITIALIZED)
         self.config = self.repo.config
         self.args = args
 

dvc/cli/parser.py

@@ -14,6 +14,7 @@ from dvc.commands import (
     config,
     daemon,
     dag,
+    data,
     data_sync,
     destroy,
     diff,
@@ -33,6 +34,7 @@ from dvc.commands import (
     move,
     params,
     plots,
+    queue,
     remote,
     remove,
     repro,
@@ -50,6 +52,7 @@ logger = logging.getLogger(__name__)
 
 COMMANDS = [
     init,
+    queue,
     get,
     get_url,
     destroy,
@@ -86,6 +89,7 @@ COMMANDS = [
     experiments,
     check_ignore,
     machine,
+    data,
 ]
 
 

dvc/commands/checkout.py

@@ -55,6 +55,10 @@ class CmdCheckout(CmdBase):
 
         if exc:
             raise exc
+
+        if self.args.relink:
+            msg = "Relinked successfully"
+            ui.write(msg)
         return 0
 
 

dvc/commands/dag.py

@@ -1,11 +1,15 @@
 import argparse
+from typing import TYPE_CHECKING
 
 from dvc.cli.command import CmdBase
 from dvc.cli.utils import append_doc_link
 from dvc.ui import ui
 
+if TYPE_CHECKING:
+    from networkx import DiGraph
 
-def _show_ascii(G):
+
+def _show_ascii(G: "DiGraph"):
     from dvc.dagascii import draw
     from dvc.repo.graph import get_pipelines
 
@@ -18,13 +22,25 @@ def _show_ascii(G):
     return "\n".join(ret)
 
 
-def _show_dot(G):
+def _quote_label(node):
+    label = str(node)
+    # Node names should not contain ":" unless they are quoted with "".
+    # See: https://github.com/pydot/pydot/issues/258.
+    if label[0] != '"' and label[-1] != '"':
+        return f'"{label}"'
+    return label
+
+
+def _show_dot(G: "DiGraph"):
     import io
 
+    import networkx as nx
     from networkx.drawing.nx_pydot import write_dot
 
     dot_file = io.StringIO()
-    write_dot(G, dot_file)
+
+    nx.relabel_nodes(G, _quote_label, copy=False)
+    write_dot(G.reverse(), dot_file)
     return dot_file.getvalue()
 
 
@@ -42,7 +58,7 @@ def _show_mermaid(G, markdown: bool = False):
         for node in nodes:
             total_nodes += 1
             node_id = f"node{total_nodes}"
-            graph += f"\n\t{node_id}[{node}]"
+            graph += f'\n\t{node_id}["{node}"]'
             node_ids[node] = node_id
         edges = sorted((str(a), str(b)) for b, a in pipeline.edges)
         for a, b in edges:
@@ -70,9 +86,9 @@ def _collect_targets(repo, target, outs):
             targets.extend([str(out) for out in stage.outs])
             continue
 
-        for out in outs_trie.itervalues(
+        for out in outs_trie.itervalues(  # noqa: B301
             prefix=repo.fs.path.parts(path)
-        ):  # noqa: B301
+        ):
             targets.extend(str(out))
 
     return targets

dvc/commands/data.py

@@ -0,0 +1,181 @@
+import argparse
+import logging
+from typing import TYPE_CHECKING
+
+from funcy import chunks, compact, log_durations
+
+from dvc.cli.command import CmdBase
+from dvc.cli.utils import append_doc_link, fix_subparsers
+from dvc.ui import ui
+from dvc.utils import colorize
+
+if TYPE_CHECKING:
+    from dvc.repo.data import Status as DataStatus
+
+
+logger = logging.getLogger(__name__)
+
+
+class CmdDataStatus(CmdBase):
+    COLORS = {
+        "not_in_cache": "red",
+        "committed": "green",
+        "uncommitted": "yellow",
+        "untracked": "cyan",
+    }
+    LABELS = {
+        "not_in_cache": "Not in cache",
+        "committed": "DVC committed changes",
+        "uncommitted": "DVC uncommitted changes",
+        "untracked": "Untracked files",
+        "unchanged": "DVC unchanged files",
+    }
+    HINTS = {
+        "not_in_cache": 'use "dvc pull <file>..." to download files',
+        "committed": "git commit the corresponding dvc files "
+        "to update the repo",
+        "uncommitted": 'use "dvc commit <file>..." to track changes',
+        "untracked": 'use "git add <file> ..." or '
+        'dvc add <file>..." to commit to git or to dvc',
+        "git_dirty": "there are {}changes not tracked by dvc, "
+        'use "git status" to see',
+    }
+
+    @staticmethod
+    def _process_status(status: "DataStatus"):
+        """Flatten stage status, and filter empty stage status contents."""
+        for stage, stage_status in status.items():
+            items = stage_status
+            if isinstance(stage_status, dict):
+                items = {
+                    file: state
+                    for state, files in stage_status.items()
+                    for file in files
+                    if state != "unknown"
+                }
+            if not items:
+                continue
+            yield stage, items
+
+    @classmethod
+    def _show_status(cls, status: "DataStatus") -> int:
+        git_info = status.pop("git")  # type: ignore[misc]
+        result = dict(cls._process_status(status))
+        if not result:
+            no_changes = "No changes"
+            if git_info.get("is_empty", False):
+                no_changes += " in an empty git repo"
+            ui.write(f"{no_changes}.")
+
+        for idx, (stage, stage_status) in enumerate(result.items()):
+            if idx:
+                ui.write()
+
+            label = cls.LABELS.get(stage, stage.capitalize() + " files")
+            header = f"{label}:"
+            color = cls.COLORS.get(stage, None)
+
+            ui.write(header)
+            if hint := cls.HINTS.get(stage):
+                ui.write(f"  ({hint})")
+
+            if isinstance(stage_status, dict):
+                items = [
+                    ": ".join([state, file])
+                    for file, state in stage_status.items()
+                ]
+            else:
+                items = stage_status
+
+            tabs = "\t".expandtabs(8)
+            for chunk in chunks(1000, items):
+                out = "\n".join(tabs + item for item in chunk)
+                ui.write(colorize(out, color))
+
+        if (hint := cls.HINTS.get("git_dirty")) and git_info.get("is_dirty"):
+            message = hint.format("other " if result else "")
+            ui.write(f"[blue]({message})[/]", styled=True)
+        return 0
+
+    def run(self) -> int:
+        with log_durations(logger.trace, "in data_status"):  # type: ignore
+            status = self.repo.data_status(
+                granular=self.args.granular,
+                untracked_files=self.args.untracked_files,
+                with_dirs=self.args.with_dirs,
+            )
+
+        if not self.args.unchanged:
+            status.pop("unchanged")  # type: ignore[misc]
+        if self.args.untracked_files == "no":
+            status.pop("untracked")
+        if self.args.json:
+            status.pop("git")  # type: ignore[misc]
+            ui.write_json(compact(status))
+            return 0
+        return self._show_status(status)
+
+
+def add_parser(subparsers, parent_parser):
+    data_parser = subparsers.add_parser(
+        "data",
+        parents=[parent_parser],
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    data_subparsers = data_parser.add_subparsers(
+        dest="cmd",
+        help="Use `dvc data CMD --help` to display command-specific help.",
+    )
+    fix_subparsers(data_subparsers)
+
+    DATA_STATUS_HELP = (
+        "Show changes between the last git commit, "
+        "the dvcfiles and the workspace."
+    )
+    data_status_parser = data_subparsers.add_parser(
+        "status",
+        parents=[parent_parser],
+        description=append_doc_link(DATA_STATUS_HELP, "data/status"),
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        help=DATA_STATUS_HELP,
+    )
+    data_status_parser.add_argument(
+        "--json",
+        action="store_true",
+        default=False,
+        help="Show output in JSON format.",
+    )
+    data_status_parser.add_argument(
+        "--show-json",
+        action="store_true",
+        default=False,
+        dest="json",
+        help=argparse.SUPPRESS,
+    )
+    data_status_parser.add_argument(
+        "--granular",
+        action="store_true",
+        default=False,
+        help="Show granular file-level info for DVC-tracked directories.",
+    )
+    data_status_parser.add_argument(
+        "--unchanged",
+        action="store_true",
+        default=False,
+        help="Show unmodified DVC-tracked files.",
+    )
+    data_status_parser.add_argument(
+        "--untracked-files",
+        choices=["no", "all"],
+        default="no",
+        const="all",
+        nargs="?",
+        help="Show untracked files.",
+    )
+    data_status_parser.add_argument(
+        "--with-dirs",
+        action="store_true",
+        default=False,
+        help=argparse.SUPPRESS,
+    )
+    data_status_parser.set_defaults(func=CmdDataStatus)

dvc/commands/data_sync.py

@@ -43,7 +43,8 @@ class CmdDataPull(CmdDataBase):
             )
             self.log_summary(stats)
         except (CheckoutError, DvcException) as exc:
-            self.log_summary(getattr(exc, "stats", {}))
+            if stats := getattr(exc, "stats", {}):
+                self.log_summary(stats)
             logger.exception("failed to pull data from the cloud")
             return 1
 

dvc/commands/experiments/__init__.py

@@ -15,6 +15,7 @@ from dvc.commands.experiments import (
     ls,
     pull,
     push,
+    queue_worker,
     remove,
     run,
     show,
@@ -30,6 +31,7 @@ SUB_COMMANDS = [
     ls,
     pull,
     push,
+    queue_worker,
     remove,
     run,
     show,

dvc/commands/experiments/branch.py

@@ -29,6 +29,6 @@ def add_parser(experiments_subparsers, parent_parser):
         "experiment", help="Experiment to be promoted."
     )
     experiments_branch_parser.add_argument(
-        "branch", help="Git branch name to use."
+        "branch", help="Name for the new Git branch"
     )
     experiments_branch_parser.set_defaults(func=CmdExperimentsBranch)

dvc/commands/experiments/gc.py

@@ -12,8 +12,8 @@ logger = logging.getLogger(__name__)
 def _raise_error_if_all_disabled(**kwargs):
     if not any(kwargs.values()):
         raise InvalidArgumentError(
-            "Either of `-w|--workspace`, `-a|--all-branches`, `-T|--all-tags` "
-            "or `--all-commits` needs to be set."
+            "Either of `-w|--workspace`, `-a|--all-branches`, `-T|--all-tags`,"
+            " `--all-commits` or `--date` needs to be set."
         )
 
 
@@ -23,6 +23,7 @@ class CmdExperimentsGC(CmdBase):
             all_branches=self.args.all_branches,
             all_tags=self.args.all_tags,
             all_commits=self.args.all_commits,
+            commit_date=self.args.commit_date,
             workspace=self.args.workspace,
         )
 
@@ -31,13 +32,21 @@ class CmdExperimentsGC(CmdBase):
         msg += "the workspace"
         if self.args.all_commits:
             msg += " and all git commits"
-        elif self.args.all_branches and self.args.all_tags:
-            msg += " and all git branches and tags"
-        elif self.args.all_branches:
-            msg += " and all git branches"
-        elif self.args.all_tags:
-            msg += " and all git tags"
+        else:
+            if self.args.all_branches and self.args.all_tags:
+                msg += " and all git branches and tags"
+            elif self.args.all_branches:
+                msg += " and all git branches"
+            elif self.args.all_tags:
+                msg += " and all git tags"
+
+            if self.args.commit_date:
+                msg += (
+                    " and all git commits before date "
+                    f"{self.args.commit_date}"
+                )
         msg += " of the current repo."
+
         if self.args.queued:
             msg += " Run queued experiments will be preserved."
         else:
@@ -54,6 +63,7 @@ class CmdExperimentsGC(CmdBase):
             all_tags=self.args.all_tags,
             all_commits=self.args.all_commits,
             workspace=self.args.workspace,
+            commit_date=self.args.commit_date,
             queued=self.args.queued,
         )
 
@@ -110,6 +120,17 @@ def add_parser(experiments_subparsers, parent_parser):
         default=False,
         help="Keep experiments derived from all Git commits.",
     )
+    experiments_gc_parser.add_argument(
+        "--date",
+        type=str,
+        dest="commit_date",
+        metavar="<YYYY-MM-DD>",
+        default=None,
+        help=(
+            "Keep experiments from the commits after (inclusive) a certain "
+            "date. Date must match the extended ISO 8601 format (YYYY-MM-DD)."
+        ),
+    )
     experiments_gc_parser.add_argument(
         "--queued",
         action="store_true",

dvc/commands/experiments/init.py

@@ -25,7 +25,6 @@ class CmdExperimentsInit(CmdBase):
     DEFAULT_METRICS = "metrics.json"
     DEFAULT_PARAMS = "params.yaml"
     PLOTS = "plots"
-    DVCLIVE = "dvclive"
     DEFAULTS = {
         "code": CODE,
         "data": DATA,
@@ -33,7 +32,6 @@ class CmdExperimentsInit(CmdBase):
         "metrics": DEFAULT_METRICS,
         "params": DEFAULT_PARAMS,
         "plots": PLOTS,
-        "live": DVCLIVE,
     }
 
     def run(self):
@@ -63,7 +61,7 @@ class CmdExperimentsInit(CmdBase):
             }
         )
 
-        initialized_stage, initialized_deps = init(
+        initialized_stage, initialized_deps, initialized_out_dirs = init(
             self.repo,
             name=self.args.name,
             type=self.args.type,
@@ -72,21 +70,32 @@ class CmdExperimentsInit(CmdBase):
             interactive=self.args.interactive,
             force=self.args.force,
         )
-        self._post_init_display(initialized_stage, initialized_deps)
+        self._post_init_display(
+            initialized_stage, initialized_deps, initialized_out_dirs
+        )
         if self.args.run:
             self.repo.experiments.run(targets=[initialized_stage.addressing])
         return 0
 
     def _post_init_display(
-        self, stage: "PipelineStage", new_deps: List["Dependency"]
+        self,
+        stage: "PipelineStage",
+        new_deps: List["Dependency"],
+        new_out_dirs: List[str],
     ) -> None:
         from dvc.utils import humanize
 
-        path_fmt = "[green]{0}[/green]".format
+        path_fmt = "[green]{}[/green]".format
         if new_deps:
             deps_paths = humanize.join(map(path_fmt, new_deps))
             ui.write(f"Creating dependencies: {deps_paths}", styled=True)
 
+        if new_out_dirs:
+            out_dirs_paths = humanize.join(map(path_fmt, new_out_dirs))
+            ui.write(
+                f"Creating output directories: {out_dirs_paths}", styled=True
+            )
+
         ui.write(
             f"Creating [b]{self.args.name}[/b] stage in [green]dvc.yaml[/]",
             styled=True,
@@ -190,12 +199,11 @@ def add_parser(experiments_subparsers, parent_parser):
     )
     experiments_init_parser.add_argument(
         "--live",
-        help="Path to log dvclive outputs for your experiments"
-        f" (default: {CmdExperimentsInit.DVCLIVE})",
+        help="Path to log dvclive outputs for your experiments",
     )
     experiments_init_parser.add_argument(
         "--type",
-        choices=["default", "dl"],
+        choices=["default", "checkpoint"],
         default="default",
         help="Select type of stage to create (default: %(default)s)",
     )

dvc/commands/experiments/ls.py

@@ -59,6 +59,6 @@ def add_parser(experiments_subparsers, parent_parser):
             "If provided, experiments from the specified Git repository "
             " will be listed instead of local ones."
         ),
-        metavar="[<git_remote>]",
+        metavar="<git_remote>",
     )
     experiments_list_parser.set_defaults(func=CmdExperimentsList)

dvc/commands/experiments/queue_worker.py

@@ -0,0 +1,25 @@
+import logging
+
+from dvc.cli.command import CmdBase
+
+logger = logging.getLogger(__name__)
+
+
+class CmdQueueWorker(CmdBase):
+    """Run the exp queue worker."""
+
+    def run(self):
+        self.repo.experiments.celery_queue.worker.start(self.args.name)
+        return 0
+
+
+def add_parser(experiments_subparsers, parent_parser):
+    QUEUE_WORKER_HELP = "Run the exp queue worker."
+    parser = experiments_subparsers.add_parser(
+        "queue-worker",
+        parents=[parent_parser],
+        description=QUEUE_WORKER_HELP,
+        add_help=False,
+    )
+    parser.add_argument("name", help="Celery worker name.")
+    parser.set_defaults(func=CmdQueueWorker)

dvc/commands/experiments/remove.py

@@ -10,23 +10,22 @@ logger = logging.getLogger(__name__)
 
 
 class CmdExperimentsRemove(CmdBase):
-    def raise_error_if_all_disabled(self):
+    def check_arguments(self):
         if not any(
             [
-                self.args.experiment,
                 self.args.all_commits,
                 self.args.rev,
                 self.args.queue,
             ]
-        ):
+        ) ^ bool(self.args.experiment):
             raise InvalidArgumentError(
                 "Either provide an `experiment` argument, or use the "
-                "`--rev` or `--all-commits` flag."
+                "`--rev` or `--all-commits` or `--queue` flag."
             )
 
     def run(self):
 
-        self.raise_error_if_all_disabled()
+        self.check_arguments()
 
         removed_list = self.repo.experiments.remove(
             exp_names=self.args.experiment,

dvc/commands/experiments/show.py

@@ -138,7 +138,7 @@ def _collect_rows(
         row_dict["Experiment"] = exp.get("name", "")
         row_dict["rev"] = name_rev
         row_dict["typ"] = typ
-        row_dict["Created"] = _format_time(
+        row_dict["Created"] = format_time(
             exp.get("timestamp"), fill_value, iso
         )
         row_dict["parent"] = parent
@@ -218,7 +218,7 @@ def _sort_exp(experiments, sort_path, sort_name, typ, reverse):
     return ret
 
 
-def _format_time(datetime_obj, fill_value=FILL_VALUE, iso=False):
+def format_time(datetime_obj, fill_value=FILL_VALUE, iso=False):
     if datetime_obj is None:
         return fill_value
 

dvc/commands/freeze.py

@@ -16,7 +16,7 @@ class CmdFreezeBase(CmdBase):
             try:
                 func(target)
             except DvcException:
-                logger.exception(f"failed to {name} '{target}'")
+                logger.exception("failed to %s '%s'", name, target)
                 ret = 1
         return ret
 

dvc/commands/gc.py

@@ -11,29 +11,46 @@ logger = logging.getLogger(__name__)
 
 class CmdGC(CmdBase):
     def run(self):
-        from dvc.repo.gc import _raise_error_if_all_disabled
+        from dvc.repo.gc import _validate_args
 
-        _raise_error_if_all_disabled(
+        _validate_args(
             all_branches=self.args.all_branches,
             all_tags=self.args.all_tags,
             all_commits=self.args.all_commits,
+            all_experiments=self.args.all_experiments,
+            commit_date=self.args.commit_date,
             workspace=self.args.workspace,
+            rev=self.args.rev,
+            num=self.args.num,
         )
 
+        if self.args.rev:
+            self.args.num = self.args.num or 1
+
         msg = "This will remove all cache except items used in "
 
         msg += "the workspace"
         if self.args.all_commits:
             msg += " and all git commits"
-        elif self.args.all_branches and self.args.all_tags:
-            msg += " and all git branches and tags"
-        elif self.args.all_branches:
-            msg += " and all git branches"
-        elif self.args.all_tags:
-            msg += " and all git tags"
-        elif self.args.all_experiments:
-            msg += " and all experiments"
+        else:
+            if self.args.all_branches and self.args.all_tags:
+                msg += " and all git branches and tags"
+            elif self.args.all_branches:
+                msg += " and all git branches"
+            elif self.args.all_tags:
+                msg += " and all git tags"
+            if self.args.commit_date:
+                msg += (
+                    " and all git commits before date "
+                    f"{self.args.commit_date}"
+                )
+            if self.args.rev:
+                msg += (
+                    f" and last {self.args.num} commits from {self.args.rev}"
+                )
 
+        if self.args.all_experiments:
+            msg += " and all experiments"
         if self.args.repos:
             msg += " of the current and the following repos:"
 
@@ -53,12 +70,15 @@ class CmdGC(CmdBase):
             all_tags=self.args.all_tags,
             all_commits=self.args.all_commits,
             all_experiments=self.args.all_experiments,
+            commit_date=self.args.commit_date,
             cloud=self.args.cloud,
             remote=self.args.remote,
             force=self.args.force,
             jobs=self.args.jobs,
             repos=self.args.repos,
             workspace=self.args.workspace,
+            rev=self.args.rev,
+            num=self.args.num,
         )
         return 0
 
@@ -83,6 +103,26 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Keep data files used in the current workspace.",
     )
+    gc_parser.add_argument(
+        "--rev",
+        type=str,
+        default=None,
+        help="Keep data files used in the specified <commit>.",
+        metavar="<commit>",
+    )
+    gc_parser.add_argument(
+        "-n",
+        "--num",
+        type=int,
+        dest="num",
+        metavar="<num>",
+        help=(
+            "Keep data files used in the last `num` commits "
+            "starting from the `--rev` <commit>. "
+            "Only used if `--rev` is also provided. "
+            "Defaults to `1`."
+        ),
+    )
     gc_parser.add_argument(
         "-a",
         "--all-branches",
@@ -104,6 +144,18 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Keep data files for all Git commits.",
     )
+    gc_parser.add_argument(
+        "--date",
+        type=str,
+        dest="commit_date",
+        metavar="<YYYY-MM-DD>",
+        default=None,
+        help=(
+            "Keep cached data referenced in the commits after ( inclusive )"
+            " a certain time. Date must match the extended ISO 8601 format "
+            "(YYYY-MM-DD)."
+        ),
+    )
     gc_parser.add_argument(
         "--all-experiments",
         action="store_true",

dvc/commands/get_url.py

@@ -18,7 +18,7 @@ class CmdGetUrl(CmdBaseNoRepo):
             Repo.get_url(self.args.url, out=self.args.out, jobs=self.args.jobs)
             return 0
         except DvcException:
-            logger.exception(f"failed to get '{self.args.url}'")
+            logger.exception("failed to get '%s'", self.args.url)
             return 1
 
 

dvc/commands/imp.py

@@ -19,6 +19,7 @@ class CmdImport(CmdBase):
                 fname=self.args.file,
                 rev=self.args.rev,
                 no_exec=self.args.no_exec,
+                no_download=self.args.no_download,
                 desc=self.args.desc,
                 jobs=self.args.jobs,
             )
@@ -69,11 +70,19 @@ def add_parser(subparsers, parent_parser):
         help="Specify name of the .dvc file this command will generate.",
         metavar="<filename>",
     )
-    import_parser.add_argument(
+    no_download_exec_group = import_parser.add_mutually_exclusive_group()
+    no_download_exec_group.add_argument(
         "--no-exec",
         action="store_true",
         default=False,
-        help="Only create .dvc file without actually downloading it.",
+        help="Only create .dvc file without actually importing target data.",
+    )
+    no_download_exec_group.add_argument(
+        "--no-download",
+        action="store_true",
+        default=False,
+        help="Create .dvc file including target data hash value(s)"
+        " but do not actually download the file(s).",
     )
     import_parser.add_argument(
         "--desc",

dvc/commands/imp_url.py

@@ -17,6 +17,7 @@ class CmdImportUrl(CmdBase):
                 out=self.args.out,
                 fname=self.args.file,
                 no_exec=self.args.no_exec,
+                no_download=self.args.no_download,
                 remote=self.args.remote,
                 to_remote=self.args.to_remote,
                 desc=self.args.desc,
@@ -66,11 +67,19 @@ def add_parser(subparsers, parent_parser):
         help="Specify name of the .dvc file this command will generate.",
         metavar="<filename>",
     ).complete = completion.DIR
-    import_parser.add_argument(
+    no_download_exec_group = import_parser.add_mutually_exclusive_group()
+    no_download_exec_group.add_argument(
         "--no-exec",
         action="store_true",
         default=False,
-        help="Only create .dvc file without actually downloading it.",
+        help="Only create .dvc file without actually importing target data.",
+    )
+    no_download_exec_group.add_argument(
+        "--no-download",
+        action="store_true",
+        default=False,
+        help="Create .dvc file including target data hash value(s)"
+        " but do not actually download the file(s).",
     )
     import_parser.add_argument(
         "--to-remote",

dvc/commands/ls/__init__.py

@@ -42,7 +42,7 @@ class CmdList(CmdBaseNoRepo):
                 ui.write("\n".join(entries))
             return 0
         except DvcException:
-            logger.exception(f"failed to list '{self.args.url}'")
+            logger.exception("failed to list '%s'", self.args.url)
             return 1
 
 

dvc/commands/plots.py

@@ -24,6 +24,60 @@ def _show_json(renderers, split=False):
     ui.write_json(result)
 
 
+def _adjust_vega_renderers(renderers):
+    from dvc.render import VERSION_FIELD
+    from dvc_render import VegaRenderer
+
+    for r in renderers:
+        if isinstance(r, VegaRenderer):
+            if _data_versions_count(r) > 1:
+                summary = _summarize_version_infos(r)
+                for dp in r.datapoints:
+                    vi = dp.pop(VERSION_FIELD, {})
+                    keys = list(vi.keys())
+                    for key in keys:
+                        if not (len(summary.get(key, set())) > 1):
+                            vi.pop(key)
+                    if vi:
+                        dp["rev"] = "::".join(vi.values())
+            else:
+                for dp in r.datapoints:
+                    dp.pop(VERSION_FIELD, {})
+
+
+def _summarize_version_infos(renderer):
+    from collections import defaultdict
+
+    from dvc.render import VERSION_FIELD
+
+    result = defaultdict(set)
+
+    for dp in renderer.datapoints:
+        for key, value in dp.get(VERSION_FIELD, {}).items():
+            result[key].add(value)
+    return dict(result)
+
+
+def _data_versions_count(renderer):
+    from itertools import product
+
+    summary = _summarize_version_infos(renderer)
+    x = product(summary.get("filename", {None}), summary.get("field", {None}))
+    return len(set(x))
+
+
+def _filter_unhandled_renderers(renderers):
+    # filtering out renderers currently unhandled by vscode extension
+    from dvc_render import VegaRenderer
+
+    def _is_json_viable(r):
+        return not (
+            isinstance(r, VegaRenderer) and _data_versions_count(r) > 1
+        )
+
+    return list(filter(_is_json_viable, renderers))
+
+
 class CmdPlots(CmdBase):
     def _func(self, *args, **kwargs):
         raise NotImplementedError
@@ -35,13 +89,30 @@ class CmdPlots(CmdBase):
         props = {p: getattr(self.args, p) for p in PLOT_PROPS}
         return {k: v for k, v in props.items() if v is not None}
 
+    def _config_files(self):
+        config_files = None
+        if self.args.from_config:
+            config_files = {self.args.from_config}
+        return config_files
+
+    def _html_template_path(self):
+        html_template_path = self.args.html_template
+        if not html_template_path:
+            html_template_path = self.repo.config.get("plots", {}).get(
+                "html_template", None
+            )
+            if html_template_path and not os.path.isabs(html_template_path):
+                html_template_path = os.path.join(
+                    self.repo.dvc_dir, html_template_path
+                )
+        return html_template_path
+
     def run(self):
         from pathlib import Path
 
+        from dvc.render.match import match_defs_renderers
         from dvc_render import render_html
 
-        from dvc.render.match import match_renderers
-
         if self.args.show_vega:
             if not self.args.targets:
                 logger.error("please specify a target for `--show-vega`")
@@ -59,9 +130,10 @@ class CmdPlots(CmdBase):
                 return 1
 
         try:
-
             plots_data = self._func(
-                targets=self.args.targets, props=self._props()
+                targets=self.args.targets,
+                props=self._props(),
+                config_files=self._config_files(),
             )
 
             if not plots_data:
@@ -70,54 +142,51 @@ class CmdPlots(CmdBase):
                     "visualization file will not be created."
                 )
 
-            out: str = self.args.out or "dvc_plots"
+            out: str = self.args.out or self.repo.config.get("plots", {}).get(
+                "out_dir", "dvc_plots"
+            )
 
             renderers_out = (
                 out if self.args.json else os.path.join(out, "static")
             )
-            renderers = match_renderers(
-                plots_data=plots_data, out=renderers_out
+
+            renderers = match_defs_renderers(
+                data=plots_data,
+                out=renderers_out,
+                templates_dir=self.repo.plots.templates_dir,
             )
+            if self.args.json:
+                renderers = _filter_unhandled_renderers(renderers)
+                _show_json(renderers, self.args.split)
+                return 0
+
+            _adjust_vega_renderers(renderers)
 
             if self.args.show_vega:
                 renderer = first(filter(lambda r: r.TYPE == "vega", renderers))
                 if renderer:
                     ui.write_json(json.loads(renderer.get_filled_template()))
                 return 0
-            if self.args.json:
-                _show_json(renderers, self.args.split)
-                return 0
-
-            html_template_path = self.args.html_template
-            if not html_template_path:
-                html_template_path = self.repo.config.get("plots", {}).get(
-                    "html_template", None
-                )
-                if html_template_path and not os.path.isabs(
-                    html_template_path
-                ):
-                    html_template_path = os.path.join(
-                        self.repo.dvc_dir, html_template_path
-                    )
 
             output_file: Path = (Path.cwd() / out).resolve() / "index.html"
 
-            render_html(
-                renderers=renderers,
-                output_file=output_file,
-                template_path=html_template_path,
-            )
+            if renderers:
+                render_html(
+                    renderers=renderers,
+                    output_file=output_file,
+                    template_path=self._html_template_path(),
+                )
 
-            ui.write(output_file.as_uri())
-            auto_open = self.repo.config["plots"].get("auto_open", False)
-            if self.args.open or auto_open:
-                if not auto_open:
-                    ui.write(
-                        "To enable auto opening, you can run:\n"
-                        "\n"
-                        "\tdvc config plots.auto_open true"
-                    )
-                return ui.open_browser(output_file)
+                ui.write(output_file.as_uri())
+                auto_open = self.repo.config["plots"].get("auto_open", False)
+                if self.args.open or auto_open:
+                    if not auto_open:
+                        ui.write(
+                            "To enable auto opening, you can run:\n"
+                            "\n"
+                            "\tdvc config plots.auto_open true"
+                        )
+                    return ui.open_browser(output_file)
 
             return 0
 
@@ -161,6 +230,8 @@ class CmdPlotsTemplates(CmdBase):
         "confusion_normalized",
         "scatter",
         "smooth",
+        "bar_horizontal_sorted",
+        "bar_horizontal",
     ]
 
     def run(self):
@@ -185,10 +256,7 @@ class CmdPlotsTemplates(CmdBase):
 
 
 def add_parser(subparsers, parent_parser):
-    PLOTS_HELP = (
-        "Commands to visualize and compare plot metrics in structured files "
-        "(JSON, YAML, CSV, TSV)."
-    )
+    PLOTS_HELP = "Commands to visualize and compare plot data."
 
     plots_parser = subparsers.add_parser(
         "plots",
@@ -204,7 +272,10 @@ def add_parser(subparsers, parent_parser):
 
     fix_subparsers(plots_subparsers)
 
-    SHOW_HELP = "Generate plots from metrics files."
+    SHOW_HELP = (
+        "Generate plots from target files or from `plots`"
+        " definitions in `dvc.yaml`."
+    )
     plots_show_parser = plots_subparsers.add_parser(
         "show",
         parents=[parent_parser],
@@ -215,9 +286,10 @@ def add_parser(subparsers, parent_parser):
     plots_show_parser.add_argument(
         "targets",
         nargs="*",
-        help="Files to visualize (supports any file, "
-        "even when not found as `plots` in `dvc.yaml`). "
-        "Shows all plots by default.",
+        help=(
+            "Plots files or plot IDs from `dvc.yaml` to visualize. "
+            "Shows all plots by default."
+        ),
     ).complete = completion.FILE
     _add_props_arguments(plots_show_parser)
     _add_output_argument(plots_show_parser)
@@ -225,8 +297,8 @@ def add_parser(subparsers, parent_parser):
     plots_show_parser.set_defaults(func=CmdPlotsShow)
 
     PLOTS_DIFF_HELP = (
-        "Show multiple versions of plot metrics "
-        "by plotting them in a single image."
+        "Show multiple versions of a plot by overlaying them "
+        "in a single image."
     )
     plots_diff_parser = plots_subparsers.add_parser(
         "diff",
@@ -239,8 +311,8 @@ def add_parser(subparsers, parent_parser):
         "--targets",
         nargs="*",
         help=(
-            "Specific plots file(s) to visualize "
-            "(even if not found as `plots` in `dvc.yaml`). "
+            "Specific plots to visualize. "
+            "Accepts any file path or plot name from `dvc.yaml` file. "
             "Shows all tracked plots by default."
         ),
         metavar="<paths>",
@@ -262,7 +334,7 @@ def add_parser(subparsers, parent_parser):
 
     PLOTS_MODIFY_HELP = (
         "Modify display properties of data-series plots "
-        "(has no effect on image-type plots)."
+        "defined in stages (has no effect on image plots)."
     )
     plots_modify_parser = plots_subparsers.add_parser(
         "modify",
@@ -272,7 +344,8 @@ def add_parser(subparsers, parent_parser):
         formatter_class=argparse.RawDescriptionHelpFormatter,
     )
     plots_modify_parser.add_argument(
-        "target", help="Metric file to set properties to"
+        "target",
+        help="Plots file to set properties for (defined at the stage level).",
     ).complete = completion.FILE
     _add_props_arguments(plots_modify_parser)
     plots_modify_parser.add_argument(
@@ -382,3 +455,9 @@ def _add_ui_arguments(parser):
         help="Custom HTML template for VEGA visualization.",
         metavar="<path>",
     )
+    parser.add_argument(
+        "--from-config",
+        default=None,
+        metavar="<path>",
+        help=argparse.SUPPRESS,
+    )

dvc/commands/queue/__init__.py

@@ -0,0 +1,34 @@
+import argparse
+
+from dvc.cli.utils import append_doc_link, fix_subparsers
+from dvc.commands.queue import kill, logs, remove, start, status, stop
+
+SUB_COMMANDS = [
+    start,
+    stop,
+    status,
+    logs,
+    remove,
+    kill,
+]
+
+
+def add_parser(subparsers, parent_parser):
+    QUEUE_HELP = "Commands to manage experiments queue."
+
+    queue_parser = subparsers.add_parser(
+        "queue",
+        parents=[parent_parser],
+        description=append_doc_link(QUEUE_HELP, "queue"),
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+        help=QUEUE_HELP,
+    )
+
+    queue_subparsers = queue_parser.add_subparsers(
+        dest="cmd",
+        help="Use `dvc queue CMD --help` to display " "command-specific help.",
+    )
+
+    fix_subparsers(queue_subparsers)
+    for cmd in SUB_COMMANDS:
+        cmd.add_parser(queue_subparsers, parent_parser)

dvc/commands/queue/kill.py

@@ -0,0 +1,34 @@
+import argparse
+import logging
+
+from dvc.cli.command import CmdBase
+from dvc.cli.utils import append_doc_link
+
+logger = logging.getLogger(__name__)
+
+
+class CmdQueueKill(CmdBase):
+    """Kill exp task in queue."""
+
+    def run(self):
+        self.repo.experiments.celery_queue.kill(revs=self.args.task)
+
+        return 0
+
+
+def add_parser(queue_subparsers, parent_parser):
+    QUEUE_KILL_HELP = "Kill actively running experiment queue tasks."
+    queue_kill_parser = queue_subparsers.add_parser(
+        "kill",
+        parents=[parent_parser],
+        description=append_doc_link(QUEUE_KILL_HELP, "queue/kill"),
+        help=QUEUE_KILL_HELP,
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    queue_kill_parser.add_argument(
+        "task",
+        nargs="*",
+        help="Tasks in queue to kill.",
+        metavar="<task>",
+    )
+    queue_kill_parser.set_defaults(func=CmdQueueKill)

dvc/commands/queue/logs.py

@@ -0,0 +1,56 @@
+import argparse
+import logging
+
+from dvc.cli.command import CmdBase
+from dvc.cli.utils import append_doc_link
+
+logger = logging.getLogger(__name__)
+
+
+class CmdQueueLogs(CmdBase):
+    """Show output logs for a queued experiment."""
+
+    def run(self):
+        self.repo.experiments.celery_queue.logs(
+            rev=self.args.task,
+            encoding=self.args.encoding,
+            follow=self.args.follow,
+        )
+
+        return 0
+
+
+def add_parser(queue_subparsers, parent_parser):
+    QUEUE_LOGS_HELP = (
+        "Show output logs for running and completed experiment queue tasks."
+    )
+    queue_logs_parser = queue_subparsers.add_parser(
+        "logs",
+        parents=[parent_parser],
+        description=append_doc_link(QUEUE_LOGS_HELP, "queue/logs"),
+        help=QUEUE_LOGS_HELP,
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    queue_logs_parser.add_argument(
+        "-e",
+        "--encoding",
+        help=(
+            "Text encoding for log output. Defaults to system locale encoding."
+        ),
+        metavar="<encoding>",
+    )
+    queue_logs_parser.add_argument(
+        "-f",
+        "--follow",
+        help=(
+            "Attach to task and follow additional live output. Only "
+            "applicable if the task is still running."
+        ),
+        action="store_true",
+    )
+    queue_logs_parser.add_argument(
+        "task",
+        help="Task to show.",
+        metavar="<task>",
+    )
+    queue_logs_parser.set_defaults(func=CmdQueueLogs)

dvc/commands/queue/remove.py

@@ -0,0 +1,95 @@
+import argparse
+import logging
+
+from dvc.cli.command import CmdBase
+from dvc.cli.utils import append_doc_link
+from dvc.exceptions import InvalidArgumentError
+from dvc.ui import ui
+
+logger = logging.getLogger(__name__)
+
+
+class CmdQueueRemove(CmdBase):
+    """Remove exp in queue."""
+
+    def check_arguments(self):
+        clear_flag = any(
+            [
+                self.args.all,
+                self.args.queued,
+                self.args.failed,
+                self.args.success,
+            ]
+        )
+        if not (clear_flag ^ bool(self.args.task)):
+            raise InvalidArgumentError(
+                "Either provide an `tasks` argument, or use the "
+                "`--all`, `--queued`, `--failed`, `--success` flag."
+            )
+
+    def run(self):
+
+        self.check_arguments()
+
+        if self.args.all:
+            self.args.queued = True
+            self.args.failed = True
+            self.args.success = True
+
+        if self.args.queued or self.args.failed or self.args.success:
+            removed_list = self.repo.experiments.celery_queue.clear(
+                success=self.args.success,
+                queued=self.args.queued,
+                failed=self.args.failed,
+            )
+        else:
+            removed_list = self.repo.experiments.celery_queue.remove(
+                revs=self.args.task,
+            )
+
+        if removed_list:
+            removed = ", ".join(removed_list)
+            ui.write(f"Removed tasks in queue: {removed}")
+        else:
+            ui.write(f"No tasks found named {self.args.task}")
+
+        return 0
+
+
+def add_parser(queue_subparsers, parent_parser):
+
+    QUEUE_REMOVE_HELP = "Remove queued and completed tasks from the queue."
+    queue_remove_parser = queue_subparsers.add_parser(
+        "remove",
+        parents=[parent_parser],
+        description=append_doc_link(QUEUE_REMOVE_HELP, "queue/remove"),
+        help=QUEUE_REMOVE_HELP,
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    queue_remove_parser.add_argument(
+        "--all",
+        action="store_true",
+        help="Remove all queued and completed tasks from the queue.",
+    )
+    queue_remove_parser.add_argument(
+        "--queued",
+        action="store_true",
+        help="Remove all queued tasks from the queue.",
+    )
+    queue_remove_parser.add_argument(
+        "--success",
+        action="store_true",
+        help="Remove all successful tasks from the queue.",
+    )
+    queue_remove_parser.add_argument(
+        "--failed",
+        action="store_true",
+        help="Remove all failed tasks from the queue.",
+    )
+    queue_remove_parser.add_argument(
+        "task",
+        nargs="*",
+        help="Tasks to remove.",
+        metavar="<task>",
+    )
+    queue_remove_parser.set_defaults(func=CmdQueueRemove)

dvc/commands/queue/start.py

@@ -0,0 +1,48 @@
+import argparse
+import logging
+
+from dvc.cli.command import CmdBase
+from dvc.cli.utils import append_doc_link
+from dvc.ui import ui
+
+logger = logging.getLogger(__name__)
+
+
+class CmdQueueStart(CmdBase):
+    """Start exp queue workers."""
+
+    def run(self):
+        started = self.repo.experiments.celery_queue.start_workers(
+            self.args.jobs
+        )
+
+        suffix = "s" if started > 1 else ""
+        ui.write(
+            f"Started '{started}' new experiments task queue worker{suffix}."
+        )
+
+        return 0
+
+
+def add_parser(queue_subparsers, parent_parser):
+
+    QUEUE_START_HELP = "Start the experiments task queue worker."
+    queue_start_parser = queue_subparsers.add_parser(
+        "start",
+        parents=[parent_parser],
+        description=append_doc_link(QUEUE_START_HELP, "queue/start"),
+        help=QUEUE_START_HELP,
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    queue_start_parser.add_argument(
+        "-j",
+        "--jobs",
+        type=int,
+        default=1,
+        help=(
+            "Maximum number of concurrent queue workers to start. "
+            "Defaults to 1."
+        ),
+        metavar="<number>",
+    )
+    queue_start_parser.set_defaults(func=CmdQueueStart)

dvc/commands/queue/status.py

@@ -0,0 +1,62 @@
+import argparse
+import logging
+from typing import List, Mapping, Optional
+
+from dvc.cli.command import CmdBase
+from dvc.cli.utils import append_doc_link
+from dvc.compare import TabularData
+from dvc.ui import ui
+
+from ..experiments.show import format_time
+
+logger = logging.getLogger(__name__)
+
+
+class CmdQueueStatus(CmdBase):
+    """Show queue task and worker status."""
+
+    def run(self):
+        result: List[
+            Mapping[str, Optional[str]]
+        ] = self.repo.experiments.celery_queue.status()
+        if result:
+            all_headers = ["Task", "Name", "Created", "Status"]
+            td = TabularData(all_headers)
+            for exp in result:
+                created = format_time(exp.get("timestamp"))
+                td.append(
+                    [
+                        exp["rev"][:7],
+                        exp.get("name", ""),
+                        created,
+                        exp["status"],
+                    ]
+                )
+            td.render()
+        else:
+            ui.write("No experiment tasks in the queue.")
+        ui.write()
+
+        worker_status = self.repo.experiments.celery_queue.worker_status()
+        active_count = len(
+            [name for name, task in worker_status.items() if task]
+        )
+        idle_count = len(worker_status) - active_count
+
+        ui.write(f"Worker status: {active_count} active, {idle_count} idle")
+
+        return 0
+
+
+def add_parser(queue_subparsers, parent_parser):
+    QUEUE_STATUS_HELP = (
+        "Show the status of experiments queue tasks and workers."
+    )
+    queue_status_parser = queue_subparsers.add_parser(
+        "status",
+        parents=[parent_parser],
+        description=append_doc_link(QUEUE_STATUS_HELP, "queue/status"),
+        help=QUEUE_STATUS_HELP,
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    queue_status_parser.set_defaults(func=CmdQueueStatus)

dvc/commands/queue/stop.py

@@ -0,0 +1,43 @@
+import argparse
+import logging
+
+from dvc.cli.command import CmdBase
+from dvc.cli.utils import append_doc_link
+from dvc.ui import ui
+
+logger = logging.getLogger(__name__)
+
+
+class CmdQueueStop(CmdBase):
+    """Stop exp queue workers."""
+
+    def run(self):
+        self.repo.experiments.celery_queue.shutdown(kill=self.args.kill)
+
+        if self.args.kill:
+            ui.write(
+                "All running tasks in the queue have been killed."
+                "Queue workers are stopping."
+            )
+        else:
+            ui.write("Queue workers will stop after running tasks finish.")
+
+        return 0
+
+
+def add_parser(queue_subparsers, parent_parser):
+
+    QUEUE_STOP_HELP = "Stop all experiments task queue workers."
+    queue_stop_parser = queue_subparsers.add_parser(
+        "stop",
+        parents=[parent_parser],
+        description=append_doc_link(QUEUE_STOP_HELP, "queue/stop"),
+        help=QUEUE_STOP_HELP,
+        formatter_class=argparse.RawDescriptionHelpFormatter,
+    )
+    queue_stop_parser.add_argument(
+        "--kill",
+        action="store_true",
+        help="Kill all running tasks before stopping the queue workers.",
+    )
+    queue_stop_parser.set_defaults(func=CmdQueueStop)

dvc/commands/repro.py

@@ -16,9 +16,7 @@ class CmdRepro(CmdBase):
         if len(stages) == 0:
             ui.write(CmdDataStatus.UP_TO_DATE_MSG)
         else:
-            ui.write(
-                "Use `dvc push` to send your updates to " "remote storage."
-            )
+            ui.write("Use `dvc push` to send your updates to remote storage.")
 
         if self.args.metrics:
             from dvc.compare import show_metrics
@@ -66,6 +64,13 @@ def add_arguments(repro_parser):
         default=False,
         help="Reproduce even if dependencies were not changed.",
     )
+    repro_parser.add_argument(
+        "-i",
+        "--interactive",
+        action="store_true",
+        default=False,
+        help="Ask for confirmation before reproducing each stage.",
+    )
     repro_parser.add_argument(
         "-s",
         "--single-item",
@@ -74,27 +79,6 @@ def add_arguments(repro_parser):
         help="Reproduce only single data item without recursive dependencies "
         "check.",
     )
-    repro_parser.add_argument(
-        "-m",
-        "--metrics",
-        action="store_true",
-        default=False,
-        help="Show metrics after reproduction.",
-    )
-    repro_parser.add_argument(
-        "--dry",
-        action="store_true",
-        default=False,
-        help="Only print the commands that would be executed without "
-        "actually executing.",
-    )
-    repro_parser.add_argument(
-        "-i",
-        "--interactive",
-        action="store_true",
-        default=False,
-        help="Ask for confirmation before reproducing each stage.",
-    )
     repro_parser.add_argument(
         "-p",
         "--pipeline",
@@ -118,11 +102,11 @@ def add_arguments(repro_parser):
         help="Reproduce all stages in the specified directory.",
     )
     repro_parser.add_argument(
-        "--force-downstream",
+        "-m",
+        "--metrics",
         action="store_true",
         default=False,
-        help="Reproduce all descendants of a changed stage even if their "
-        "direct dependencies didn't change.",
+        help="Show metrics after reproduction.",
     )
     repro_parser.add_argument(
         "--downstream",
@@ -130,6 +114,13 @@ def add_arguments(repro_parser):
         default=False,
         help="Start from the specified stages when reproducing pipelines.",
     )
+    repro_parser.add_argument(
+        "--force-downstream",
+        action="store_true",
+        default=False,
+        help="Reproduce all descendants of a changed stage even if their "
+        "direct dependencies didn't change.",
+    )
     repro_parser.add_argument(
         "--pull",
         action="store_true",
@@ -139,6 +130,13 @@ def add_arguments(repro_parser):
             "from the run-cache."
         ),
     )
+    repro_parser.add_argument(
+        "--dry",
+        action="store_true",
+        default=False,
+        help="Only print the commands that would be executed without "
+        "actually executing.",
+    )
 
 
 def add_parser(subparsers, parent_parser):
@@ -156,13 +154,10 @@ def add_parser(subparsers, parent_parser):
     add_arguments(repro_parser)
     # repro only args
     repro_parser.add_argument(
-        "--no-run-cache",
+        "--glob",
         action="store_true",
         default=False,
-        help=(
-            "Execute stage commands even if they have already been run with "
-            "the same command/dependencies/outputs/etc before."
-        ),
+        help="Allows targets containing shell-style wildcards.",
     )
     repro_parser.add_argument(
         "--no-commit",
@@ -171,9 +166,12 @@ def add_parser(subparsers, parent_parser):
         help="Don't put files/directories into cache.",
     )
     repro_parser.add_argument(
-        "--glob",
+        "--no-run-cache",
         action="store_true",
         default=False,
-        help="Allows targets containing shell-style wildcards.",
+        help=(
+            "Execute stage commands even if they have already been run with "
+            "the same command/dependencies/outputs/etc before."
+        ),
     )
     repro_parser.set_defaults(func=CmdRepro)

dvc/commands/run.py

@@ -69,12 +69,25 @@ def add_parser(subparsers, parent_parser):
     run_parser.add_argument(
         "--file", metavar="<filename>", help=argparse.SUPPRESS
     )
+    run_parser.add_argument(
+        "--single-stage",
+        action="store_true",
+        default=False,
+        help=argparse.SUPPRESS,
+    )
+    _add_common_args(run_parser)
     run_parser.add_argument(
         "--no-exec",
         action="store_true",
         default=False,
         help="Only create dvc.yaml without actually running it.",
     )
+    run_parser.add_argument(
+        "--no-commit",
+        action="store_true",
+        default=False,
+        help="Don't put files/directories into cache.",
+    )
     run_parser.add_argument(
         "--no-run-cache",
         action="store_true",
@@ -84,17 +97,4 @@ def add_parser(subparsers, parent_parser):
             "with the same command/dependencies/outputs/etc before."
         ),
     )
-    run_parser.add_argument(
-        "--no-commit",
-        action="store_true",
-        default=False,
-        help="Don't put files/directories into cache.",
-    )
-    run_parser.add_argument(
-        "--single-stage",
-        action="store_true",
-        default=False,
-        help=argparse.SUPPRESS,
-    )
-    _add_common_args(run_parser)
     run_parser.set_defaults(func=CmdRun)

dvc/commands/stage.py

@@ -137,6 +137,13 @@ class CmdStageAdd(CmdBase):
 
 
 def _add_common_args(parser):
+    parser.add_argument(
+        "-f",
+        "--force",
+        action="store_true",
+        default=False,
+        help="Overwrite existing stage",
+    )
     parser.add_argument(
         "-d",
         "--deps",
@@ -145,6 +152,14 @@ def _add_common_args(parser):
         help="Declare dependencies for reproducible cmd.",
         metavar="<path>",
     ).complete = completion.FILE
+    parser.add_argument(
+        "-p",
+        "--params",
+        action="append",
+        default=[],
+        help="Declare parameter to use as additional dependency.",
+        metavar="[<filename>:]<params_list>",
+    ).complete = completion.FILE
     parser.add_argument(
         "-o",
         "--outs",
@@ -163,13 +178,36 @@ def _add_common_args(parser):
         metavar="<filename>",
     ).complete = completion.FILE
     parser.add_argument(
-        "-p",
-        "--params",
+        "-c",
+        "--checkpoints",
         action="append",
         default=[],
-        help="Declare parameter to use as additional dependency.",
-        metavar="[<filename>:]<params_list>",
+        help="Declare checkpoint output file or directory for 'dvc exp run'. "
+        "Not compatible with 'dvc repro'.",
+        metavar="<filename>",
     ).complete = completion.FILE
+    parser.add_argument(
+        "--external",
+        action="store_true",
+        default=False,
+        help="Allow outputs that are outside of the DVC repository.",
+    )
+    parser.add_argument(
+        "--outs-persist",
+        action="append",
+        default=[],
+        help="Declare output file or directory that will not be "
+        "removed upon repro.",
+        metavar="<filename>",
+    )
+    parser.add_argument(
+        "--outs-persist-no-cache",
+        action="append",
+        default=[],
+        help="Declare output file or directory that will not be "
+        "removed upon repro (do not put into DVC cache).",
+        metavar="<filename>",
+    )
     parser.add_argument(
         "-m",
         "--metrics",
@@ -226,50 +264,12 @@ def _add_common_args(parser):
         help="Directory within your repo to run your command in.",
         metavar="<path>",
     )
-    parser.add_argument(
-        "-f",
-        "--force",
-        action="store_true",
-        default=False,
-        help="Overwrite existing stage",
-    )
-    parser.add_argument(
-        "--outs-persist",
-        action="append",
-        default=[],
-        help="Declare output file or directory that will not be "
-        "removed upon repro.",
-        metavar="<filename>",
-    )
-    parser.add_argument(
-        "--outs-persist-no-cache",
-        action="append",
-        default=[],
-        help="Declare output file or directory that will not be "
-        "removed upon repro (do not put into DVC cache).",
-        metavar="<filename>",
-    )
-    parser.add_argument(
-        "-c",
-        "--checkpoints",
-        action="append",
-        default=[],
-        help="Declare checkpoint output file or directory for 'dvc exp run'. "
-        "Not compatible with 'dvc repro'.",
-        metavar="<filename>",
-    ).complete = completion.FILE
     parser.add_argument(
         "--always-changed",
         action="store_true",
         default=False,
         help="Always consider this DVC-file as changed.",
     )
-    parser.add_argument(
-        "--external",
-        action="store_true",
-        default=False,
-        help="Allow outputs that are outside of the DVC repository.",
-    )
     parser.add_argument(
         "--desc",
         type=str,

dvc/commands/update.py

@@ -18,6 +18,7 @@ class CmdUpdate(CmdBase):
                 rev=self.args.rev,
                 recursive=self.args.recursive,
                 to_remote=self.args.to_remote,
+                no_download=self.args.no_download,
                 remote=self.args.remote,
                 jobs=self.args.jobs,
             )
@@ -28,7 +29,10 @@ class CmdUpdate(CmdBase):
 
 
 def add_parser(subparsers, parent_parser):
-    UPDATE_HELP = "Update data artifacts imported from other DVC repositories."
+    UPDATE_HELP = (
+        "Update data artifact imported (via dvc import or dvc import-url) "
+        "from an external DVC repository or URL."
+    )
     update_parser = subparsers.add_parser(
         "update",
         parents=[parent_parser],
@@ -52,6 +56,13 @@ def add_parser(subparsers, parent_parser):
         default=False,
         help="Update all stages in the specified directory.",
     )
+    update_parser.add_argument(
+        "--no-download",
+        action="store_true",
+        default=False,
+        help="Update .dvc file hash value(s)"
+        " but do not download the file(s).",
+    )
     update_parser.add_argument(
         "--to-remote",
         action="store_true",

dvc/config.py

@@ -86,9 +86,11 @@ class Config(dict):
     def __init__(
         self, dvc_dir=None, validate=True, fs=None, config=None
     ):  # pylint: disable=super-init-not-called
-        from dvc.fs.local import LocalFileSystem
+        from dvc.fs import LocalFileSystem
 
         self.dvc_dir = dvc_dir
+        self.wfs = LocalFileSystem()
+        self.fs = fs or self.wfs
 
         if not dvc_dir:
             try:
@@ -98,10 +100,7 @@ class Config(dict):
             except NotDvcRepoError:
                 self.dvc_dir = None
         else:
-            self.dvc_dir = os.path.abspath(os.path.realpath(dvc_dir))
-
-        self.wfs = LocalFileSystem(url=self.dvc_dir)
-        self.fs = fs or self.wfs
+            self.dvc_dir = self.fs.path.abspath(self.fs.path.realpath(dvc_dir))
 
         self.load(validate=validate, config=config)
 
@@ -124,8 +123,8 @@ class Config(dict):
         }
 
         if self.dvc_dir is not None:
-            files["repo"] = os.path.join(self.dvc_dir, self.CONFIG)
-            files["local"] = os.path.join(self.dvc_dir, self.CONFIG_LOCAL)
+            files["repo"] = self.fs.path.join(self.dvc_dir, self.CONFIG)
+            files["local"] = self.fs.path.join(self.dvc_dir, self.CONFIG_LOCAL)
 
         return files
 
@@ -188,7 +187,7 @@ class Config(dict):
         filename = self.files[level]
         fs = self._get_fs(level)
 
-        logger.debug(f"Writing '{filename}'.")
+        logger.debug("Writing '%s'.", filename)
 
         fs.makedirs(os.path.dirname(filename))
 
@@ -228,7 +227,7 @@ class Config(dict):
 
     @staticmethod
     def _to_relpath(conf_dir, path):
-        from dvc.fs.local import localfs
+        from dvc.fs import localfs
         from dvc.utils import relpath
 
         from .config_schema import RelPath

dvc/config_schema.py

@@ -152,6 +152,8 @@ SCHEMA = {
                     "ssl_verify": Any(Bool, str),
                     "sse": str,
                     "sse_kms_key_id": str,
+                    "sse_customer_algorithm": str,
+                    "sse_customer_key": str,
                     "acl": str,
                     "grant_read": str,
                     "grant_read_acp": str,
@@ -166,6 +168,7 @@ SCHEMA = {
                 "gs": {
                     "projectname": str,
                     "credentialpath": str,
+                    "endpointurl": str,
                     Optional("verify", default=False): Bool,
                     **REMOTE_COMMON,
                 },
@@ -217,6 +220,7 @@ SCHEMA = {
                     **REMOTE_COMMON,
                 },
                 "gdrive": {
+                    "profile": str,
                     "gdrive_use_service_account": Bool,
                     "gdrive_client_id": str,
                     "gdrive_client_secret": str,
@@ -224,6 +228,7 @@ SCHEMA = {
                     "gdrive_service_account_user_email": str,
                     "gdrive_service_account_json_file_path": str,
                     Optional("gdrive_trash_only", default=False): Bool,
+                    Optional("gdrive_acknowledge_abuse", default=False): Bool,
                     Optional("verify", default=True): Bool,
                     **REMOTE_COMMON,
                 },
@@ -269,6 +274,7 @@ SCHEMA = {
     "plots": {
         "html_template": str,
         Optional("auto_open", default=False): Bool,
+        "out_dir": str,
     },
     "exp": {
         "code": str,
@@ -279,4 +285,8 @@ SCHEMA = {
         "plots": str,
         "live": str,
     },
+    "parsing": {
+        "bool": All(Lower, Choices("store_true", "boolean_optional")),
+        "list": All(Lower, Choices("nargs", "append")),
+    },
 }

dvc/daemon.py

@@ -6,6 +6,7 @@ import os
 import platform
 import sys
 from subprocess import Popen
+from typing import List
 
 from dvc.env import DVC_DAEMON
 from dvc.utils import fix_env, is_binary
@@ -81,7 +82,7 @@ def _spawn_posix(cmd, env):
 
 
 def _spawn(cmd, env):
-    logger.debug(f"Trying to spawn '{cmd}'")
+    logger.debug("Trying to spawn '%s'", cmd)
 
     if os.name == "nt":
         _spawn_windows(cmd, env)
@@ -90,7 +91,7 @@ def _spawn(cmd, env):
     else:
         raise NotImplementedError
 
-    logger.debug(f"Spawned '{cmd}'")
+    logger.debug("Spawned '%s'", cmd)
 
 
 def daemon(args):
@@ -99,12 +100,14 @@ def daemon(args):
     Args:
         args (list): list of arguments to append to `dvc daemon` command.
     """
+    daemonize(["daemon", "-q"] + args)
+
+
+def daemonize(cmd: List[str]):
     if os.environ.get(DVC_DAEMON):
         logger.debug("skipping launching a new daemon.")
         return
 
-    cmd = ["daemon", "-q"] + args
-
     env = fix_env()
     if not is_binary():
         file_path = os.path.abspath(inspect.stack()[0][1])

dvc/data/__init__.py

@@ -1,33 +0,0 @@
-import logging
-from typing import TYPE_CHECKING, Iterator, Union
-
-from .tree import Tree
-
-if TYPE_CHECKING:
-    from dvc.hash_info import HashInfo
-    from dvc.objects.db import ObjectDB
-    from dvc.objects.file import HashFile
-
-logger = logging.getLogger(__name__)
-
-
-def check(odb: "ObjectDB", obj: "HashFile", **kwargs):
-    if isinstance(obj, Tree):
-        for _, _, oid in obj:
-            odb.check(oid, **kwargs)
-
-    odb.check(obj.hash_info, **kwargs)
-
-
-def load(odb: "ObjectDB", hash_info: "HashInfo") -> "HashFile":
-    if hash_info.isdir:
-        return Tree.load(odb, hash_info)
-    return odb.get(hash_info)
-
-
-def iterobjs(
-    obj: Union["Tree", "HashFile"]
-) -> Iterator[Union["Tree", "HashFile"]]:
-    if isinstance(obj, Tree):
-        yield from (entry_obj for _, entry_obj in obj)
-    yield obj

dvc/data/_progress.py

@@ -1,33 +0,0 @@
-from dvc.progress import Tqdm
-
-
-class QueryingProgress(Tqdm):
-    def __init__(self, iterable=None, total=None, name=None, phase="Querying"):
-        msg_part = "cache in " + f"'{name}'" if name else "remote cache"
-        msg_fmt = "{phase} " + msg_part
-
-        self._estimating_msg = msg_fmt.format(phase="Estimating size of")
-        self._listing_msg = msg_fmt.format(phase="Querying")
-        self.desc = desc = msg_fmt.format(phase=phase)
-        super().__init__(
-            iterable=iterable,
-            desc=desc,
-            total=total,
-            unit="files",
-            unit_scale=False,
-            bar_format=self.BAR_FMT_DEFAULT,
-        )
-
-    def callback(self, phase, *args):
-        total = args[0] if args else self.total
-        completed = args[1] if len(args) > 1 else self.n
-        desc = self.desc
-        if phase == "estimating":
-            desc = self._estimating_msg
-        elif phase == "querying":
-            desc = self._listing_msg
-
-        if desc != self.desc:
-            # let Tqdm take care of when to update
-            self.set_description_str(desc, refresh=False)
-        self.update_to(completed, total)

dvc/data/checkout.py

@@ -1,265 +0,0 @@
-import logging
-from itertools import chain
-
-from dvc import prompt
-from dvc.exceptions import CacheLinkError, CheckoutError, ConfirmRemoveError
-from dvc.fs.utils import test_links, transfer
-from dvc.ignore import DvcIgnoreFilter
-from dvc.types import Optional
-
-from .diff import ROOT
-from .diff import diff as odiff
-from .slow_link_detection import slow_link_guard  # type: ignore[attr-defined]
-from .stage import stage
-
-logger = logging.getLogger(__name__)
-
-
-def _remove(fs_path, fs, in_cache, force=False):
-    if not fs.exists(fs_path):
-        return
-
-    if force:
-        fs.remove(fs_path)
-        return
-
-    if not in_cache:
-        msg = (
-            f"file/directory '{fs_path}' is going to be removed. "
-            "Are you sure you want to proceed?"
-        )
-
-        if not prompt.confirm(msg):
-            raise ConfirmRemoveError(fs_path)
-
-    fs.remove(fs_path)
-
-
-def _relink(link, cache, cache_info, fs, path, in_cache, force):
-    _remove(path, fs, in_cache, force=force)
-    link(cache, cache_info, fs, path)
-    # NOTE: Depending on a file system (e.g. on NTFS), `_remove` might reset
-    # read-only permissions in order to delete a hardlink to protected object,
-    # which will also reset it for the object itself, making it unprotected,
-    # so we need to protect it back.
-    cache.protect(cache_info)
-
-
-def _checkout_file(
-    link,
-    fs_path,
-    fs,
-    change,
-    cache,
-    force,
-    progress_callback=None,
-    relink=False,
-    state=None,
-):
-    """The file is changed we need to checkout a new copy"""
-    modified = False
-
-    cache_fs_path = cache.hash_to_path(change.new.oid.value)
-    if change.old.oid:
-        if relink:
-            if fs.iscopy(fs_path) and cache.cache_types[0] == "copy":
-                cache.unprotect(fs_path)
-            else:
-                _relink(
-                    link,
-                    cache,
-                    cache_fs_path,
-                    fs,
-                    fs_path,
-                    change.old.in_cache,
-                    force=force,
-                )
-        else:
-            modified = True
-            _relink(
-                link,
-                cache,
-                cache_fs_path,
-                fs,
-                fs_path,
-                change.old.in_cache,
-                force=force,
-            )
-    else:
-        link(cache, cache_fs_path, fs, fs_path)
-        modified = True
-
-    if state:
-        state.save(fs_path, fs, change.new.oid)
-
-    if progress_callback:
-        progress_callback(fs_path)
-
-    return modified
-
-
-def _diff(
-    fs_path,
-    fs,
-    obj,
-    cache,
-    relink=False,
-    dvcignore: Optional[DvcIgnoreFilter] = None,
-):
-    old = None
-    try:
-        _, _, old = stage(
-            cache,
-            fs_path,
-            fs,
-            obj.hash_info.name if obj else cache.fs.PARAM_CHECKSUM,
-            dry_run=True,
-            dvcignore=dvcignore,
-        )
-    except FileNotFoundError:
-        pass
-
-    diff = odiff(old, obj, cache)
-
-    if relink:
-        diff.modified.extend(diff.unchanged)
-
-    return diff
-
-
-class Link:
-    def __init__(self, links):
-        self._links = links
-
-    @slow_link_guard
-    def __call__(self, cache, from_path, to_fs, to_path):
-        if to_fs.exists(to_path):
-            to_fs.remove(to_path)  # broken symlink
-
-        cache.makedirs(cache.fs.path.parent(to_path))
-        try:
-            transfer(cache.fs, from_path, to_fs, to_path, links=self._links)
-        except FileNotFoundError as exc:
-            raise CheckoutError([to_path]) from exc
-        except OSError as exc:
-            raise CacheLinkError([to_path]) from exc
-
-
-def _checkout(
-    diff,
-    fs_path,
-    fs,
-    cache,
-    force=False,
-    progress_callback=None,
-    relink=False,
-    state=None,
-):
-    if not diff:
-        return
-
-    links = test_links(cache.cache_types, cache.fs, cache.fs_path, fs, fs_path)
-    if not links:
-        raise CacheLinkError([fs_path])
-    link = Link(links)
-
-    for change in diff.deleted:
-        entry_path = (
-            fs.path.join(fs_path, *change.old.key)
-            if change.old.key != ROOT
-            else fs_path
-        )
-        _remove(entry_path, fs, change.old.in_cache, force=force)
-
-    failed = []
-    for change in chain(diff.added, diff.modified):
-        entry_path = (
-            fs.path.join(fs_path, *change.new.key)
-            if change.new.key != ROOT
-            else fs_path
-        )
-        if change.new.oid.isdir:
-            fs.makedirs(entry_path)
-            continue
-
-        try:
-            _checkout_file(
-                link,
-                entry_path,
-                fs,
-                change,
-                cache,
-                force,
-                progress_callback,
-                relink,
-                state=state,
-            )
-        except CheckoutError as exc:
-            failed.extend(exc.target_infos)
-
-    if failed:
-        raise CheckoutError(failed)
-
-
-def checkout(
-    fs_path,
-    fs,
-    obj,
-    cache,
-    force=False,
-    progress_callback=None,
-    relink=False,
-    quiet=False,
-    dvcignore: Optional[DvcIgnoreFilter] = None,
-    state=None,
-):
-    # if scheme(fs_path) not in ["local", cache.fs.scheme]:
-    #    raise NotImplementedError
-
-    diff = _diff(
-        fs_path,
-        fs,
-        obj,
-        cache,
-        relink=relink,
-        dvcignore=dvcignore,
-    )
-
-    failed = []
-    if not obj:
-        if not quiet:
-            logger.warning(
-                "No file hash info found for '%s'. It won't be created.",
-                fs_path,
-            )
-        failed.append(fs_path)
-    elif not diff:
-        logger.trace("Data '%s' didn't change.", fs_path)  # type: ignore
-
-    try:
-        _checkout(
-            diff,
-            fs_path,
-            fs,
-            cache,
-            force=force,
-            progress_callback=progress_callback,
-            relink=relink,
-            state=state,
-        )
-    except CheckoutError as exc:
-        failed.extend(exc.target_infos)
-
-    if diff and state:
-        state.save_link(fs_path, fs)
-        if not failed:
-            state.save(fs_path, fs, obj.hash_info)
-
-    if failed or not diff:
-        if progress_callback and obj:
-            progress_callback(fs_path, len(obj))
-        if failed:
-            raise CheckoutError(failed)
-        return
-
-    return bool(diff) and not relink

dvc/data/db/index.py

@@ -1,149 +0,0 @@
-import logging
-import os
-from abc import ABC, abstractmethod
-from typing import TYPE_CHECKING, Iterable, Set
-
-from dvc.objects.errors import ObjectDBError
-from dvc.utils.decorators import with_diskcache
-
-if TYPE_CHECKING:
-    from dvc.types import StrPath
-
-logger = logging.getLogger(__name__)
-
-
-class ObjectDBIndexBase(ABC):
-    @abstractmethod
-    def __init__(
-        self,
-        tmp_dir: "StrPath",
-        name: str,
-    ):
-        pass
-
-    @abstractmethod
-    def __iter__(self):
-        pass
-
-    @abstractmethod
-    def __contains__(self, hash_):
-        pass
-
-    def hashes(self):
-        return iter(self)
-
-    @abstractmethod
-    def dir_hashes(self):
-        pass
-
-    @abstractmethod
-    def clear(self):
-        pass
-
-    @abstractmethod
-    def update(self, dir_hashes: Iterable[str], file_hashes: Iterable[str]):
-        pass
-
-    @abstractmethod
-    def intersection(self, hashes: Set[str]):
-        pass
-
-
-class ObjectDBIndexNoop(ObjectDBIndexBase):
-    """No-op class for ODBs which are not indexed."""
-
-    def __init__(
-        self,
-        tmp_dir: "StrPath",
-        name: str,
-    ):  # pylint: disable=super-init-not-called
-        pass
-
-    def __iter__(self):
-        return iter([])
-
-    def __contains__(self, hash_):
-        return False
-
-    def dir_hashes(self):
-        return []
-
-    def clear(self):
-        pass
-
-    def update(self, dir_hashes: Iterable[str], file_hashes: Iterable[str]):
-        pass
-
-    def intersection(self, hashes: Set[str]):
-        return []
-
-
-class ObjectDBIndex(ObjectDBIndexBase):
-    """Class for indexing hashes in an ODB."""
-
-    INDEX_SUFFIX = ".idx"
-    INDEX_DIR = "index"
-
-    def __init__(
-        self,
-        tmp_dir: "StrPath",
-        name: str,
-    ):  # pylint: disable=super-init-not-called
-        from diskcache import Cache, Index
-
-        from dvc.fs.local import LocalFileSystem
-        from dvc.utils.fs import makedirs
-
-        self.index_dir = os.path.join(tmp_dir, self.INDEX_DIR, name)
-        makedirs(self.index_dir, exist_ok=True)
-        self.fs = LocalFileSystem()
-        self.index = Index.fromcache(
-            Cache(
-                self.index_dir,
-                disk_pickle_protocol=4,
-                eviction_policy="none",
-            )
-        )
-
-    @with_diskcache(name="index")
-    def __iter__(self):
-        return iter(self.index)
-
-    @with_diskcache(name="index")
-    def __contains__(self, hash_):
-        return hash_ in self.index
-
-    @with_diskcache(name="index")
-    def dir_hashes(self):
-        """Iterate over .dir hashes stored in the index."""
-        yield from (hash_ for hash_, is_dir in self.index.items() if is_dir)
-
-    @with_diskcache(name="index")
-    def clear(self):
-        """Clear this index (to force re-indexing later)."""
-        from diskcache import Timeout
-
-        try:
-            self.index.clear()
-        except Timeout as exc:
-            raise ObjectDBError("Failed to clear ODB index") from exc
-
-    @with_diskcache(name="index")
-    def update(self, dir_hashes: Iterable[str], file_hashes: Iterable[str]):
-        """Update this index, adding the specified hashes."""
-        from diskcache import Timeout
-
-        try:
-            with self.index.transact():
-                for hash_ in dir_hashes:
-                    self.index[hash_] = True
-            with self.index.transact():
-                for hash_ in file_hashes:
-                    self.index[hash_] = False
-        except Timeout as exc:
-            raise ObjectDBError("Failed to update ODB index") from exc
-
-    @with_diskcache(name="index")
-    def intersection(self, hashes: Set[str]):
-        """Iterate over values from `hashes` which exist in the index."""
-        yield from hashes.intersection(self.index.keys())

dvc/data/db/local.py

@@ -1,166 +0,0 @@
-import logging
-import os
-import stat
-from functools import partial
-
-from funcy import cached_property
-from shortuuid import uuid
-
-from dvc.hash_info import HashInfo
-from dvc.objects.db import ObjectDB, noop, wrap_iter
-from dvc.objects.errors import ObjectFormatError
-from dvc.utils import relpath
-from dvc.utils.fs import copyfile, remove, umask, walk_files
-
-logger = logging.getLogger(__name__)
-
-
-class LocalObjectDB(ObjectDB):
-    DEFAULT_CACHE_TYPES = ["reflink", "copy"]
-    CACHE_MODE = 0o444
-    UNPACKED_DIR_SUFFIX = ".unpacked"
-
-    def __init__(self, fs, fs_path, **config):
-        super().__init__(fs, fs_path, **config)
-        self.cache_dir = fs_path
-
-        shared = config.get("shared")
-        if shared:
-            self._file_mode = 0o664
-            self._dir_mode = 0o2775
-        else:
-            self._file_mode = 0o666 & ~umask
-            self._dir_mode = 0o777 & ~umask
-
-    @property
-    def cache_dir(self):
-        return self.fs_path if self.fs_path else None
-
-    @cache_dir.setter
-    def cache_dir(self, value):
-        self.fs_path = value
-
-    @cached_property
-    def cache_path(self):
-        return os.path.abspath(self.cache_dir)
-
-    def move(self, from_info, to_info):
-        super().move(from_info, to_info)
-        os.chmod(to_info, self._file_mode)
-
-    def makedirs(self, fs_path):
-        from dvc.utils.fs import makedirs
-
-        makedirs(fs_path, exist_ok=True, mode=self._dir_mode)
-
-    def hash_to_path(self, hash_):
-        # NOTE: `self.cache_path` is already normalized so we can simply use
-        # `os.sep` instead of `os.path.join`. This results in this helper
-        # being ~5.5 times faster.
-        return f"{self.cache_path}{os.sep}{hash_[0:2]}{os.sep}{hash_[2:]}"
-
-    def hashes_exist(
-        self, hashes, jobs=None, progress=noop
-    ):  # pylint: disable=unused-argument
-        ret = []
-        progress = partial(progress, "querying", len(hashes))
-
-        for hash_ in wrap_iter(hashes, progress):
-            hash_info = HashInfo(self.fs.PARAM_CHECKSUM, hash_)
-            try:
-                self.check(hash_info)
-                ret.append(hash_)
-            except (FileNotFoundError, ObjectFormatError):
-                pass
-
-        return ret
-
-    def _list_paths(self, prefix=None):
-        assert self.fs_path is not None
-        if prefix:
-            fs_path = self.fs.path.join(self.fs_path, prefix[:2])
-            if not self.fs.exists(fs_path):
-                return
-        else:
-            fs_path = self.fs_path
-
-        # NOTE: use utils.fs walk_files since fs.walk_files will not follow
-        # symlinks
-        yield from walk_files(fs_path)
-
-    def _remove_unpacked_dir(self, hash_):
-        hash_fs_path = self.hash_to_path(hash_)
-        fs_path = self.fs.path.with_name(
-            hash_fs_path,
-            self.fs.path.name(hash_fs_path) + self.UNPACKED_DIR_SUFFIX,
-        )
-        self.fs.remove(fs_path)
-
-    def _unprotect_file(self, path):
-        if self.fs.is_symlink(path) or self.fs.is_hardlink(path):
-            logger.debug("Unprotecting '%s'", path)
-            tmp = os.path.join(os.path.dirname(path), "." + uuid())
-
-            # The operations order is important here - if some application
-            # would access the file during the process of copyfile then it
-            # would get only the part of file. So, at first, the file should be
-            # copied with the temporary name, and then original file should be
-            # replaced by new.
-            copyfile(path, tmp, name=f"Unprotecting '{relpath(path)}'")
-            remove(path)
-            os.rename(tmp, path)
-
-        else:
-            logger.debug(
-                "Skipping copying for '%s', since it is not "
-                "a symlink or a hardlink.",
-                path,
-            )
-
-        os.chmod(path, self._file_mode)
-
-    def _unprotect_dir(self, path):
-        for fname in self.fs.find(path):
-            self._unprotect_file(fname)
-
-    def unprotect(self, fs_path):
-        if not os.path.exists(fs_path):
-            from dvc.exceptions import DvcException
-
-            raise DvcException(
-                f"can't unprotect non-existing data '{fs_path}'"
-            )
-
-        if os.path.isdir(fs_path):
-            self._unprotect_dir(fs_path)
-        else:
-            self._unprotect_file(fs_path)
-
-    def protect(self, fs_path):
-        try:
-            os.chmod(fs_path, self.CACHE_MODE)
-        except OSError:
-            # NOTE: not being able to protect cache file is not fatal, it
-            # might happen on funky filesystems (e.g. Samba, see #5255),
-            # read-only filesystems or in a shared cache scenario.
-            logger.trace("failed to protect '%s'", fs_path, exc_info=True)
-
-    def is_protected(self, fs_path):
-        try:
-            mode = os.stat(fs_path).st_mode
-        except FileNotFoundError:
-            return False
-
-        return stat.S_IMODE(mode) == self.CACHE_MODE
-
-    def set_exec(self, fs_path):
-        mode = os.stat(fs_path).st_mode | stat.S_IEXEC
-        try:
-            os.chmod(fs_path, mode)
-        except OSError:
-            logger.trace(
-                "failed to chmod '%s' '%s'",
-                oct(mode),
-                fs_path,
-                exc_info=True,
-            )

dvc/data/db/reference.py

@@ -1,84 +0,0 @@
-import logging
-import os
-from typing import TYPE_CHECKING, Dict
-
-from dvc.objects.db import ObjectDB
-from dvc.objects.errors import ObjectFormatError
-from dvc.scheme import Schemes
-
-from ..reference import ReferenceHashFile
-
-if TYPE_CHECKING:
-    from dvc.fs.base import AnyFSPath, FileSystem
-    from dvc.hash_info import HashInfo
-
-logger = logging.getLogger(__name__)
-
-
-class ReferenceObjectDB(ObjectDB):
-    """Reference ODB.
-
-    File objects are stored as ReferenceHashFiles which reference paths outside
-    of the staging ODB fs. Tree objects are stored natively.
-    """
-
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        self._fs_cache: Dict[tuple, "FileSystem"] = {}
-        self._obj_cache: Dict["HashInfo", "ReferenceHashFile"] = {}
-
-    def get(self, hash_info: "HashInfo"):
-        if hash_info.isdir:
-            return super().get(hash_info)
-        try:
-            return self._obj_cache[hash_info]
-        except KeyError:
-            pass
-        fs_path = self.hash_to_path(hash_info.value)
-        try:
-            ref_file = ReferenceHashFile.from_bytes(
-                self.fs.cat_file(fs_path),
-                fs_cache=self._fs_cache,
-            )
-        except OSError:
-            raise FileNotFoundError
-        try:
-            ref_file.check(self, check_hash=False)
-        except ObjectFormatError:
-            self.fs.remove(fs_path)
-            raise
-        self._obj_cache[hash_info] = ref_file
-        return ref_file
-
-    def _add_file(
-        self,
-        from_fs: "FileSystem",
-        from_info: "AnyFSPath",
-        to_info: "AnyFSPath",
-        hash_info: "HashInfo",
-        hardlink: bool = False,
-    ):
-        self.makedirs(self.fs.path.parent(to_info))
-        if hash_info.isdir:
-            return super()._add_file(
-                from_fs,
-                from_info,
-                to_info,
-                hash_info,
-                hardlink=hardlink,
-            )
-        ref_file = ReferenceHashFile(from_info, from_fs, hash_info)
-        self._obj_cache[hash_info] = ref_file
-        try:
-            self.fs.pipe_file(to_info, ref_file.to_bytes())
-        except OSError as exc:
-            if isinstance(exc, FileExistsError) or (
-                os.name == "nt"
-                and exc.__context__
-                and isinstance(exc.__context__, FileExistsError)
-            ):
-                logger.debug("'%s' file already exists, skipping", to_info)
-            else:
-                raise
-        if from_fs.scheme != Schemes.LOCAL:
-            self._fs_cache[ReferenceHashFile.config_tuple(from_fs)] = from_fs

dvc/data/diff.py

@@ -1,133 +0,0 @@
-from dataclasses import dataclass, field
-from typing import TYPE_CHECKING, List, Optional, Tuple
-
-if TYPE_CHECKING:
-    from dvc.hash_info import HashInfo
-    from dvc.objects.file import HashFile
-
-ADD = "add"
-MODIFY = "modify"
-DELETE = "delete"
-UNCHANGED = "unchanged"
-
-
-@dataclass
-class TreeEntry:
-    in_cache: bool
-    key: Tuple[str]
-    oid: Optional["HashInfo"] = field(default=None)
-
-    def __bool__(self):
-        return bool(self.oid)
-
-    def __eq__(self, other):
-        if not isinstance(other, TreeEntry):
-            return False
-
-        if self.key != other.key:
-            return False
-
-        return self.oid == other.oid
-
-
-@dataclass
-class Change:
-    old: TreeEntry
-    new: TreeEntry
-
-    @property
-    def typ(self):
-        if not self.old and not self.new:
-            return UNCHANGED
-
-        if self.old and not self.new:
-            return DELETE
-
-        if not self.old and self.new:
-            return ADD
-
-        if self.old != self.new:
-            return MODIFY
-
-        return UNCHANGED
-
-    def __bool__(self):
-        return self.typ != UNCHANGED
-
-
-@dataclass
-class DiffResult:
-    added: List[Change] = field(default_factory=list, compare=True)
-    modified: List[Change] = field(default_factory=list, compare=True)
-    deleted: List[Change] = field(default_factory=list, compare=True)
-    unchanged: List[Change] = field(default_factory=list, compare=True)
-
-    def __bool__(self):
-        return bool(self.added or self.modified or self.deleted)
-
-
-ROOT = ("",)
-
-
-def diff(
-    old: Optional["HashFile"], new: Optional["HashFile"], cache
-) -> DiffResult:
-    from .tree import Tree
-
-    if old is None and new is None:
-        return DiffResult()
-
-    def _get_keys(obj):
-        if not obj:
-            return []
-        return [ROOT] + (
-            [key for key, _, _ in obj] if isinstance(obj, Tree) else []
-        )
-
-    old_keys = set(_get_keys(old))
-    new_keys = set(_get_keys(new))
-
-    def _get_oid(obj, key):
-        if not obj or key == ROOT:
-            return obj.hash_info if obj else None
-
-        entry_obj = obj.get(cache, key)
-        return entry_obj.hash_info if entry_obj else None
-
-    def _in_cache(oid, cache):
-        from dvc.objects.errors import ObjectFormatError
-
-        if not oid:
-            return False
-
-        try:
-            cache.check(oid)
-            return True
-        except (FileNotFoundError, ObjectFormatError):
-            return False
-
-    ret = DiffResult()
-    for key in old_keys | new_keys:
-        old_oid = _get_oid(old, key)
-        new_oid = _get_oid(new, key)
-
-        change = Change(
-            old=TreeEntry(_in_cache(old_oid, cache), key, old_oid),
-            new=TreeEntry(_in_cache(new_oid, cache), key, new_oid),
-        )
-
-        if change.typ == ADD:
-            ret.added.append(change)
-        elif change.typ == MODIFY:
-            ret.modified.append(change)
-        elif change.typ == DELETE:
-            ret.deleted.append(change)
-        else:
-            assert change.typ == UNCHANGED
-            if not change.new.in_cache and not (
-                change.new.oid and change.new.oid.isdir
-            ):
-                ret.modified.append(change)
-            else:
-                ret.unchanged.append(change)
-    return ret

dvc/data/gc.py

@@ -1,40 +0,0 @@
-def gc(odb, used, jobs=None, cache_odb=None, shallow=True):
-    from dvc.data.tree import Tree
-    from dvc.objects.errors import ObjectDBPermissionError
-
-    from ._progress import QueryingProgress
-
-    if odb.read_only:
-        raise ObjectDBPermissionError("Cannot gc read-only ODB")
-    if not cache_odb:
-        cache_odb = odb
-    used_hashes = set()
-    for hash_info in used:
-        used_hashes.add(hash_info.value)
-        if hash_info.isdir and not shallow:
-            tree = Tree.load(cache_odb, hash_info)
-            used_hashes.update(
-                entry_obj.hash_info.value for _, entry_obj in tree
-            )
-
-    def _is_dir_hash(_hash):
-        from dvc.hash_info import HASH_DIR_SUFFIX
-
-        return _hash.endswith(HASH_DIR_SUFFIX)
-
-    removed = False
-    # hashes must be sorted to ensure we always remove .dir files first
-
-    hashes = QueryingProgress(odb.all(jobs), name=odb.fs_path)
-    for hash_ in sorted(hashes, key=_is_dir_hash, reverse=True):
-        if hash_ in used_hashes:
-            continue
-        fs_path = odb.hash_to_path(hash_)
-        if _is_dir_hash(hash_):
-            # backward compatibility
-            # pylint: disable=protected-access
-            odb._remove_unpacked_dir(hash_)
-        odb.fs.remove(fs_path)
-        removed = True
-
-    return removed

dvc/data/meta.py

@@ -1,47 +0,0 @@
-from collections import OrderedDict
-from dataclasses import dataclass, field
-from typing import TYPE_CHECKING, Optional
-
-if TYPE_CHECKING:
-    from dvc.objects.db import ObjectDB
-    from dvc.objects.file import HashFile
-
-
-@dataclass
-class Meta:
-    PARAM_SIZE = "size"
-    PARAM_NFILES = "nfiles"
-    PARAM_ISEXEC = "isexec"
-
-    size: Optional[int] = field(default=None)
-    nfiles: Optional[int] = field(default=None)
-    isexec: Optional[bool] = field(default=False)
-
-    obj: Optional["HashFile"] = field(default=None)
-    odb: Optional["ObjectDB"] = field(default=None)
-    remote: Optional[str] = field(default=None)
-
-    @classmethod
-    def from_dict(cls, d):
-        if not d:
-            return cls()
-
-        size = d.pop(cls.PARAM_SIZE, None)
-        nfiles = d.pop(cls.PARAM_NFILES, None)
-        isexec = d.pop(cls.PARAM_ISEXEC, False)
-
-        return cls(size=size, nfiles=nfiles, isexec=isexec)
-
-    def to_dict(self):
-        ret = OrderedDict()
-
-        if self.size is not None:
-            ret[self.PARAM_SIZE] = self.size
-
-        if self.nfiles is not None:
-            ret[self.PARAM_NFILES] = self.nfiles
-
-        if self.isexec:
-            ret[self.PARAM_ISEXEC] = self.isexec
-
-        return ret

dvc/data/reference.py

@@ -1,113 +0,0 @@
-import errno
-import logging
-import os
-import pickle
-from typing import TYPE_CHECKING, Optional
-
-from dvc.objects.errors import ObjectFormatError
-from dvc.objects.file import HashFile
-
-if TYPE_CHECKING:
-    from dvc.fs.base import AnyFSPath, FileSystem
-    from dvc.hash_info import HashInfo
-    from dvc.objects.db import ObjectDB
-
-logger = logging.getLogger(__name__)
-
-
-class ReferenceHashFile(HashFile):
-    PARAM_PATH = "path"
-    PARAM_HASH = "hash"
-    PARAM_CHECKSUM = "checksum"
-    PARAM_FS_CONFIG = "fs_config"
-
-    def __init__(
-        self,
-        fs_path: "AnyFSPath",
-        fs: "FileSystem",
-        hash_info: "HashInfo",
-        checksum: Optional[str] = None,
-        **kwargs,
-    ):
-        super().__init__(fs_path, fs, hash_info, **kwargs)
-        self.checksum = checksum or fs.checksum(fs_path)
-
-    def __str__(self):
-        return f"ref object {self.hash_info} -> {self.fs_path}"
-
-    def check(self, odb: "ObjectDB", check_hash: bool = True):
-        assert self.fs
-        if not self.fs.exists(self.fs_path):
-            raise FileNotFoundError(
-                errno.ENOENT, os.strerror(errno.ENOENT), self.fs_path
-            )
-        if self.checksum != self._get_checksum():
-            raise ObjectFormatError(f"{self} is changed")
-        if check_hash:
-            self._check_hash(odb)
-
-    def _get_checksum(self) -> str:
-        assert self.fs
-        return self.fs.checksum(self.fs_path)
-
-    def to_bytes(self):
-        # NOTE: dumping reference FS's this way is insecure, as the
-        # fully parsed remote FS config will include credentials
-        #
-        # ReferenceHashFiles should currently only be serialized in
-        # memory and not to disk
-        fs_path = self.fs_path
-        dict_ = {
-            self.PARAM_PATH: fs_path,
-            self.PARAM_HASH: self.hash_info,
-            self.PARAM_CHECKSUM: self.checksum,
-            self.PARAM_FS_CONFIG: self.config_tuple(self.fs),
-        }
-        try:
-            return pickle.dumps(dict_)
-        except pickle.PickleError as exc:
-            raise ObjectFormatError(f"Could not pickle {self}") from exc
-
-    @classmethod
-    def from_bytes(cls, data: bytes, fs_cache: Optional[dict] = None):
-        from dvc.fs import get_fs_cls
-        from dvc.fs.repo import RepoFileSystem, _RepoFileSystem
-
-        try:
-            dict_ = pickle.loads(data)
-        except pickle.PickleError as exc:
-            raise ObjectFormatError("ReferenceHashFile is corrupted") from exc
-
-        try:
-            fs_path = dict_[cls.PARAM_PATH]
-            hash_info = dict_[cls.PARAM_HASH]
-        except KeyError as exc:
-            raise ObjectFormatError("ReferenceHashFile is corrupted") from exc
-
-        scheme, config_pairs = dict_.get(cls.PARAM_FS_CONFIG)
-        fs = fs_cache.get((scheme, config_pairs)) if fs_cache else None
-        if not fs:
-            config = dict(config_pairs)
-            if _RepoFileSystem.PARAM_REPO_URL in config:
-                fs_cls = RepoFileSystem
-            else:
-                fs_cls = get_fs_cls(config, scheme=scheme)
-            fs = fs_cls(**config)
-        return ReferenceHashFile(
-            fs_path,
-            fs,
-            hash_info,
-            checksum=dict_.get(cls.PARAM_CHECKSUM),
-        )
-
-    @staticmethod
-    def config_tuple(fs: "FileSystem"):
-        return (
-            fs.scheme,
-            tuple(
-                (key, value)
-                for key, value in sorted(
-                    fs.config.items(), key=lambda item: item[0]
-                )
-            ),
-        )

dvc/data/slow_link_detection.py

@@ -1,45 +0,0 @@
-# type: ignore
-import logging
-import sys
-import time
-from functools import wraps
-
-import colorama
-
-logger = logging.getLogger(__name__)
-this = sys.modules[__name__]
-
-this.timeout_seconds = 10.0
-this.already_displayed = False
-this.message = (
-    "You can cut execution time considerably by using a different "
-    "`cache.type` configuration.\n"
-    "See {blue}https://dvc.org/doc/commands-reference/config#cache{reset} "
-    "for more information.\n"
-    "To disable this message, run:\n"
-    "`dvc config cache.slow_link_warning false`".format(
-        blue=colorama.Fore.BLUE, reset=colorama.Fore.RESET
-    )
-)
-
-
-def slow_link_guard(f):
-    @wraps(f)
-    def wrapper(self, odb, *args, **kwargs):
-        if this.already_displayed:
-            return f(self, odb, *args, **kwargs)
-
-        if not odb.slow_link_warning or odb.cache_types:
-            return f(self, odb, *args, **kwargs)
-
-        start = time.time()
-        result = f(self, odb, *args, **kwargs)
-        delta = time.time() - start
-
-        if delta >= this.timeout_seconds:
-            logger.warning(this.message)
-            this.already_displayed = True
-
-        return result
-
-    return wrapper

dvc/data/stage.py

@@ -1,352 +0,0 @@
-import hashlib
-import logging
-import os
-from concurrent.futures import ThreadPoolExecutor
-from functools import partial
-from typing import TYPE_CHECKING, Dict, Optional, Tuple
-
-from dvc.exceptions import DvcIgnoreInCollectedDirError
-from dvc.hash_info import HashInfo
-from dvc.ignore import DvcIgnore
-from dvc.objects.file import HashFile
-from dvc.progress import Tqdm
-from dvc.utils import file_md5, is_exec
-
-from .db.reference import ReferenceObjectDB
-from .meta import Meta
-
-if TYPE_CHECKING:
-    from dvc.fs.base import AnyFSPath, FileSystem
-    from dvc.objects.db import ObjectDB
-
-logger = logging.getLogger(__name__)
-
-
-_STAGING_MEMFS_PATH = "dvc-staging"
-
-
-def _upload_file(from_fs_path, fs, odb, upload_odb, callback=None):
-    from dvc.fs._callback import FsspecCallback
-    from dvc.utils import tmp_fname
-    from dvc.utils.stream import HashedStreamReader
-
-    fs_path = upload_odb.fs.path
-    tmp_info = fs_path.join(upload_odb.fs_path, tmp_fname())
-    with fs.open(from_fs_path, mode="rb") as stream:
-        stream = HashedStreamReader(stream)
-        size = fs.size(from_fs_path)
-        with FsspecCallback.as_tqdm_callback(
-            callback,
-            desc=fs_path.name(from_fs_path),
-            bytes=True,
-            total=size,
-        ) as cb:
-            upload_odb.fs.upload(stream, tmp_info, size=size, callback=cb)
-
-    odb.add(tmp_info, upload_odb.fs, stream.hash_info)
-    meta = Meta(size=size)
-    return from_fs_path, meta, odb.get(stream.hash_info)
-
-
-def _adapt_info(info, scheme):
-    if scheme == "s3" and "ETag" in info:
-        info["etag"] = info["ETag"].strip('"')
-    elif scheme == "gs" and "etag" in info:
-        import base64
-
-        info["etag"] = base64.b64decode(info["etag"]).hex()
-    elif scheme.startswith("http") and (
-        "ETag" in info or "Content-MD5" in info
-    ):
-        info["checksum"] = info.get("ETag") or info.get("Content-MD5")
-    return info
-
-
-def _get_file_hash(fs_path, fs, name):
-    info = _adapt_info(fs.info(fs_path), fs.scheme)
-
-    if name in info:
-        assert not info[name].endswith(".dir")
-        hash_value = info[name]
-    elif hasattr(fs, name):
-        func = getattr(fs, name)
-        hash_value = func(fs_path)
-    elif name == "md5":
-        hash_value = file_md5(fs_path, fs)
-    else:
-        raise NotImplementedError
-
-    meta = Meta(size=info["size"], isexec=is_exec(info.get("mode", 0)))
-    hash_info = HashInfo(name, hash_value)
-    return meta, hash_info
-
-
-def get_file_hash(fs_path, fs, name, state=None):
-    if state:
-        meta, hash_info = state.get(  # pylint: disable=assignment-from-none
-            fs_path, fs
-        )
-        if hash_info:
-            return meta, hash_info
-
-    meta, hash_info = _get_file_hash(fs_path, fs, name)
-
-    if state:
-        assert ".dir" not in hash_info.value
-        state.save(fs_path, fs, hash_info)
-
-    return meta, hash_info
-
-
-def _stage_file(fs_path, fs, name, odb=None, upload_odb=None, dry_run=False):
-    state = odb.state if odb else None
-    meta, hash_info = get_file_hash(fs_path, fs, name, state=state)
-    if upload_odb and not dry_run:
-        assert odb and name == "md5"
-        return _upload_file(fs_path, fs, odb, upload_odb)
-
-    if dry_run:
-        obj = HashFile(fs_path, fs, hash_info)
-    else:
-        odb.add(fs_path, fs, hash_info, hardlink=False)
-        obj = odb.get(hash_info)
-
-    return fs_path, meta, obj
-
-
-def _build_objects(
-    fs_path,
-    fs,
-    name,
-    dvcignore=None,
-    jobs=None,
-    no_progress_bar=False,
-    **kwargs,
-):
-    if dvcignore:
-        walk_iterator = dvcignore.find(fs, fs_path)
-    else:
-        walk_iterator = fs.find(fs_path)
-    with Tqdm(
-        unit="md5",
-        desc="Computing file/dir hashes (only done once)",
-        disable=no_progress_bar,
-    ) as pbar:
-        worker = pbar.wrap_fn(
-            partial(
-                _stage_file,
-                fs=fs,
-                name=name,
-                **kwargs,
-            )
-        )
-        with ThreadPoolExecutor(
-            max_workers=jobs if jobs is not None else fs.hash_jobs
-        ) as executor:
-            yield from executor.map(worker, walk_iterator)
-
-
-def _iter_objects(fs_path, fs, name, **kwargs):
-    yield from _build_objects(fs_path, fs, name, **kwargs)
-
-
-def _build_tree(fs_path, fs, name, **kwargs):
-    from .tree import Tree
-
-    tree_meta = Meta(size=0, nfiles=0)
-    tree = Tree(None, None, None)
-    for file_fs_path, meta, obj in _iter_objects(fs_path, fs, name, **kwargs):
-        if DvcIgnore.DVCIGNORE_FILE == fs.path.name(file_fs_path):
-            raise DvcIgnoreInCollectedDirError(fs.path.parent(file_fs_path))
-
-        # NOTE: this is lossy transformation:
-        #   "hey\there" -> "hey/there"
-        #   "hey/there" -> "hey/there"
-        # The latter is fine filename on Windows, which
-        # will transform to dir/file on back transform.
-        #
-        # Yes, this is a BUG, as long as we permit "/" in
-        # filenames on Windows and "\" on Unix
-
-        key = fs.path.relparts(file_fs_path, fs_path)
-        assert key
-        tree.add(key, meta, obj.hash_info)
-
-        tree_meta.size += meta.size or 0
-        tree_meta.nfiles += 1
-
-    return tree_meta, tree
-
-
-def _stage_tree(fs_path, fs, fs_info, name, odb=None, **kwargs):
-    from .tree import Tree
-
-    value = fs_info.get(name)
-    if odb and value:
-        hash_info = HashInfo(name, value)
-        try:
-            tree = Tree.load(odb, hash_info)
-            return Meta(nfiles=len(tree)), tree
-        except FileNotFoundError:
-            pass
-
-    meta, tree = _build_tree(fs_path, fs, name, odb=odb, **kwargs)
-    state = odb.state if odb and odb.state else None
-    hash_info = None
-    if state:
-        _, hash_info = state.get(  # pylint: disable=assignment-from-none
-            fs_path, fs
-        )
-    tree.digest(hash_info=hash_info)
-    odb.add(tree.fs_path, tree.fs, tree.hash_info, hardlink=False)
-    raw = odb.get(tree.hash_info)
-    # cleanup unneeded memfs tmpfile and return tree based on the
-    # ODB fs/path
-    if odb.fs != tree.fs:
-        tree.fs.remove(tree.fs_path)
-    tree.fs = raw.fs
-    tree.fs_path = raw.fs_path
-    return meta, tree
-
-
-_url_cache: Dict[str, str] = {}
-
-
-def _make_staging_url(
-    fs: "FileSystem", odb: "ObjectDB", fs_path: Optional[str]
-):
-    from dvc.scheme import Schemes
-
-    url = f"{Schemes.MEMORY}://{_STAGING_MEMFS_PATH}"
-
-    if fs_path is not None:
-        if odb.fs.scheme == Schemes.LOCAL:
-            fs_path = os.path.abspath(fs_path)
-
-        if fs_path not in _url_cache:
-            _url_cache[fs_path] = hashlib.sha256(
-                fs_path.encode("utf-8")
-            ).hexdigest()
-
-        url = fs.path.join(url, _url_cache[fs_path])
-
-    return url
-
-
-def _get_staging(odb: "ObjectDB") -> "ObjectDB":
-    """Return an ODB that can be used for staging objects.
-
-    Staging will be a reference ODB stored in the the global memfs.
-    """
-
-    from dvc.fs.memory import MemoryFileSystem
-
-    fs = MemoryFileSystem()
-    fs_path = _make_staging_url(fs, odb, odb.fs_path)
-    state = odb.state
-    return ReferenceObjectDB(fs, fs_path, state=state)
-
-
-def _load_from_state(odb, staging, fs_path, fs, name):
-    from dvc.objects.errors import ObjectFormatError
-
-    from . import check, load
-    from .tree import Tree
-
-    state = odb.state
-    meta, hash_info = state.get(fs_path, fs)
-    if hash_info:
-        for odb_ in (odb, staging):
-            if odb_.exists(hash_info):
-                try:
-                    obj = load(odb_, hash_info)
-                    check(odb_, obj, check_hash=False)
-                    if isinstance(obj, Tree):
-                        meta.nfiles = len(obj)
-                    assert obj.hash_info.name == name
-                    return odb_, meta, obj
-                except (ObjectFormatError, FileNotFoundError):
-                    pass
-    raise FileNotFoundError
-
-
-def _stage_external_tree_info(odb, tree, name):
-    # NOTE: used only for external outputs. Initial reasoning was to be
-    # able to validate .dir files right in the workspace (e.g. check s3
-    # etag), but could be dropped for manual validation with regular md5,
-    # that would be universal for all clouds.
-    assert odb and name != "md5"
-
-    odb.add(tree.fs_path, tree.fs, tree.hash_info)
-    raw = odb.get(tree.hash_info)
-    _, hash_info = get_file_hash(raw.fs_path, raw.fs, name, state=odb.state)
-    tree.fs_path = raw.fs_path
-    tree.fs = raw.fs
-    tree.hash_info.name = hash_info.name
-    tree.hash_info.value = hash_info.value
-    if not tree.hash_info.value.endswith(".dir"):
-        tree.hash_info.value += ".dir"
-    return tree
-
-
-def stage(
-    odb: "ObjectDB",
-    fs_path: "AnyFSPath",
-    fs: "FileSystem",
-    name: str,
-    upload: bool = False,
-    dry_run: bool = False,
-    **kwargs,
-) -> Tuple["ObjectDB", "Meta", "HashFile"]:
-    """Stage (prepare) objects from the given path for addition to an ODB.
-
-    Returns at tuple of (staging_odb, object) where addition to the ODB can
-    be completed by transferring the object from staging to the dest ODB.
-
-    If dry_run is True, object hashes will be computed and returned, but file
-    objects themselves will not be added to the staging ODB (i.e. the resulting
-    file objects cannot transferred from staging to another ODB).
-
-    If upload is True, files will be uploaded to a temporary path on the dest
-    ODB filesystem, and staged objects will reference the uploaded path rather
-    than the original source path.
-    """
-    assert fs_path
-    # assert scheme(fs_path) == fs.scheme
-
-    details = fs.info(fs_path)
-    staging = _get_staging(odb)
-    if odb:
-        try:
-            return _load_from_state(odb, staging, fs_path, fs, name)
-        except FileNotFoundError:
-            pass
-
-    if details["type"] == "directory":
-        meta, obj = _stage_tree(
-            fs_path,
-            fs,
-            details,
-            name,
-            odb=staging,
-            upload_odb=odb if upload else None,
-            dry_run=dry_run,
-            **kwargs,
-        )
-        logger.debug("staged tree '%s'", obj)
-        if name != "md5":
-            obj = _stage_external_tree_info(odb, obj, name)
-    else:
-        _, meta, obj = _stage_file(
-            fs_path,
-            fs,
-            name,
-            odb=staging,
-            upload_odb=odb if upload else None,
-            dry_run=dry_run,
-        )
-
-    if odb and odb.state and obj.hash_info:
-        odb.state.save(fs_path, fs, obj.hash_info)
-
-    return staging, meta, obj

dvc/data/status.py

@@ -1,207 +0,0 @@
-import logging
-from typing import TYPE_CHECKING, Dict, Iterable, NamedTuple, Optional, Set
-
-from dvc.hash_info import HashInfo
-from dvc.scheme import Schemes
-
-from .tree import Tree
-
-if TYPE_CHECKING:
-    from dvc.objects.db import ObjectDB
-    from dvc.objects.file import HashFile
-
-    from .db.index import ObjectDBIndexBase
-
-logger = logging.getLogger(__name__)
-
-
-class StatusResult(NamedTuple):
-    exists: Set["HashInfo"]
-    missing: Set["HashInfo"]
-
-
-class CompareStatusResult(NamedTuple):
-    ok: Set["HashInfo"]
-    missing: Set["HashInfo"]
-    new: Set["HashInfo"]
-    deleted: Set["HashInfo"]
-
-
-def _indexed_dir_hashes(odb, index, dir_objs, name, cache_odb, jobs=None):
-    # Validate our index by verifying all indexed .dir hashes
-    # still exist on the remote
-    from ._progress import QueryingProgress
-
-    dir_hashes = set(dir_objs.keys())
-    indexed_dirs = set(index.dir_hashes())
-    indexed_dir_exists = set()
-    if indexed_dirs:
-        hashes = QueryingProgress(
-            odb.list_hashes_exists(indexed_dirs, jobs=jobs),
-            total=len(indexed_dirs),
-        )
-        indexed_dir_exists.update(hashes)
-        missing_dirs = indexed_dirs.difference(indexed_dir_exists)
-        if missing_dirs:
-            logger.debug(
-                "Remote cache missing indexed .dir hashes '%s', "
-                "clearing remote index",
-                ", ".join(missing_dirs),
-            )
-            index.clear()
-
-    # Check if non-indexed (new) dir hashes exist on remote
-    dir_exists = dir_hashes.intersection(indexed_dir_exists)
-    dir_missing = dir_hashes - dir_exists
-    dir_exists.update(
-        QueryingProgress(
-            odb.list_hashes_exists(dir_missing, jobs=jobs),
-            total=len(dir_missing),
-        )
-    )
-
-    # If .dir hash exists in the ODB, assume directory contents
-    # also exists
-    for dir_hash in dir_exists:
-        tree = dir_objs.get(dir_hash)
-        if not tree:
-            try:
-                tree = Tree.load(cache_odb, HashInfo(name, dir_hash))
-            except FileNotFoundError:
-                continue
-        file_hashes = [oid.value for _, _, oid in tree]
-        if dir_hash not in index:
-            logger.debug(
-                "Indexing new .dir '%s' with '%s' nested files",
-                dir_hash,
-                len(file_hashes),
-            )
-            index.update([dir_hash], file_hashes)
-        yield from file_hashes
-        yield tree.hash_info.value
-
-
-def status(
-    odb: "ObjectDB",
-    obj_ids: Iterable["HashInfo"],
-    name: Optional[str] = None,
-    index: Optional["ObjectDBIndexBase"] = None,
-    cache_odb: Optional["ObjectDB"] = None,
-    shallow: bool = True,
-    jobs: Optional[int] = None,
-    **kwargs,
-) -> "StatusResult":
-    """Return status of whether or not the specified objects exist odb.
-
-    If cache_odb is set, trees will be loaded from cache_odb instead of odb
-    when needed.
-
-    Status is returned as a tuple of:
-        exists: objs that exist in odb
-        missing: objs that do not exist in ODB
-    """
-    logger.debug("Preparing to collect status from '%s'", odb.fs_path)
-    if not name:
-        name = odb.fs.PARAM_CHECKSUM
-
-    if cache_odb is None:
-        cache_odb = odb
-
-    hash_infos: Dict[str, "HashInfo"] = {}
-    dir_objs: Dict[str, Optional["HashFile"]] = {}
-    for hash_info in obj_ids:
-        assert hash_info.value
-        if hash_info.isdir:
-            if shallow:
-                tree = None
-            else:
-                tree = Tree.load(cache_odb, hash_info)
-                for _, _, oid in tree:
-                    assert oid and oid.value
-                    hash_infos[oid.value] = oid
-            if index:
-                dir_objs[hash_info.value] = tree
-        hash_infos[hash_info.value] = hash_info
-
-    if odb.fs.scheme == Schemes.MEMORY:
-        # assume memfs staged objects already exist
-        return StatusResult(set(hash_infos.values()), set())
-
-    hashes: Set[str] = set(hash_infos.keys())
-    exists: Set[str] = set()
-
-    logger.debug("Collecting status from '%s'", odb.fs_path)
-    if index and hashes:
-        if dir_objs:
-            exists = hashes.intersection(
-                _indexed_dir_hashes(
-                    odb, index, dir_objs, name, cache_odb, jobs=jobs
-                )
-            )
-            hashes.difference_update(exists)
-        if hashes:
-            exists.update(index.intersection(hashes))
-            hashes.difference_update(exists)
-
-    if hashes:
-        from ._progress import QueryingProgress
-
-        with QueryingProgress(phase="Checking", name=odb.fs_path) as pbar:
-            exists.update(
-                odb.hashes_exist(hashes, jobs=jobs, progress=pbar.callback)
-            )
-    return StatusResult(
-        {hash_infos[hash_] for hash_ in exists},
-        {hash_infos[hash_] for hash_ in (hashes - exists)},
-    )
-
-
-def compare_status(
-    src: "ObjectDB",
-    dest: "ObjectDB",
-    obj_ids: Iterable["HashInfo"],
-    log_missing: bool = True,
-    check_deleted: bool = True,
-    src_index: Optional["ObjectDBIndexBase"] = None,
-    dest_index: Optional["ObjectDBIndexBase"] = None,
-    jobs: Optional[int] = None,
-    **kwargs,
-) -> "CompareStatusResult":
-    """Compare status for the specified objects between two ODBs.
-
-    Status is returned as a tuple of:
-        ok: hashes that exist in both src and dest
-        missing: hashes that do not exist in neither src nor dest
-        new: hashes that only exist in src
-        deleted: hashes that only exist in dest
-    """
-    if "cache_odb" not in kwargs:
-        kwargs["cache_odb"] = src
-    dest_exists, dest_missing = status(
-        dest, obj_ids, index=dest_index, jobs=jobs, **kwargs
-    )
-    # for transfer operations we can skip src status check when all objects
-    # already exist in dest
-    if dest_missing or check_deleted:
-        src_exists, src_missing = status(
-            src, obj_ids, index=src_index, jobs=jobs, **kwargs
-        )
-    else:
-        src_exists = dest_exists
-        src_missing = set()
-    result = CompareStatusResult(
-        src_exists & dest_exists,
-        src_missing & dest_missing,
-        src_exists - dest_exists,
-        dest_exists - src_exists,
-    )
-    if log_missing and result.missing:
-        missing_desc = "\n".join(
-            f"name: {hash_info.obj_name}, {hash_info}"
-            for hash_info in result.missing
-        )
-        logger.warning(
-            "Some of the cache files do not exist neither locally "
-            f"nor on remote. Missing cache files:\n{missing_desc}"
-        )
-    return result

dvc/data/transfer.py

@@ -1,179 +0,0 @@
-import errno
-import logging
-from functools import partial, wraps
-from typing import TYPE_CHECKING, Any, Callable, Iterable, Optional
-
-from funcy import split
-
-from dvc.progress import Tqdm
-from dvc.utils.threadpool import ThreadPoolExecutor
-
-if TYPE_CHECKING:
-    from dvc.hash_info import HashInfo
-    from dvc.objects.db import ObjectDB
-
-    from .db.index import ObjectDBIndexBase
-    from .tree import Tree
-
-logger = logging.getLogger(__name__)
-
-
-def _log_exceptions(func):
-    @wraps(func)
-    def wrapper(fs_path, *args, **kwargs):
-        try:
-            func(fs_path, *args, **kwargs)
-            return 0
-        except Exception as exc:  # pylint: disable=broad-except
-            # NOTE: this means we ran out of file descriptors and there is no
-            # reason to try to proceed, as we will hit this error anyways.
-            # pylint: disable=no-member
-            if isinstance(exc, OSError) and exc.errno == errno.EMFILE:
-                raise
-
-            logger.exception("failed to transfer '%s'", fs_path)
-            return 1
-
-    return wrapper
-
-
-def find_tree_by_obj_id(
-    odbs: Iterable[Optional["ObjectDB"]], obj_id: "HashInfo"
-) -> Optional["Tree"]:
-    from dvc.objects.errors import ObjectFormatError
-
-    from .tree import Tree
-
-    for odb in odbs:
-        if odb is not None:
-            try:
-                return Tree.load(odb, obj_id)
-            except (FileNotFoundError, ObjectFormatError):
-                pass
-    return None
-
-
-def _do_transfer(
-    src: "ObjectDB",
-    dest: "ObjectDB",
-    obj_ids: Iterable["HashInfo"],
-    missing_ids: Iterable["HashInfo"],
-    processor: Callable,
-    src_index: Optional["ObjectDBIndexBase"] = None,
-    dest_index: Optional["ObjectDBIndexBase"] = None,
-    cache_odb: Optional["ObjectDB"] = None,
-    **kwargs: Any,
-):
-    from dvc.exceptions import FileTransferError
-
-    dir_ids, file_ids = split(lambda hash_info: hash_info.isdir, obj_ids)
-    total_fails = 0
-    succeeded_dir_objs = []
-    all_file_ids = set(file_ids)
-
-    for dir_hash in dir_ids:
-        dir_obj = find_tree_by_obj_id([cache_odb, src], dir_hash)
-        assert dir_obj
-
-        entry_ids = {oid for _, _, oid in dir_obj}
-        bound_file_ids = all_file_ids & entry_ids
-        all_file_ids -= entry_ids
-
-        dir_fails = sum(processor(bound_file_ids))
-        if dir_fails:
-            logger.debug(
-                "failed to upload full contents of '%s', "
-                "aborting .dir file upload",
-                dir_obj.name,
-            )
-            logger.error(
-                "failed to upload '%s' to '%s'",
-                src.get(dir_obj.hash_info).fs_path,
-                dest.get(dir_obj.hash_info).fs_path,
-            )
-            total_fails += dir_fails + 1
-        elif entry_ids.intersection(missing_ids):
-            # if for some reason a file contained in this dir is
-            # missing both locally and in the remote, we want to
-            # push whatever file content we have, but should not
-            # push .dir file
-            logger.debug(
-                "directory '%s' contains missing files,"
-                "skipping .dir file upload",
-                dir_obj.name,
-            )
-        else:
-            is_dir_failed = sum(processor([dir_obj.hash_info]))
-            total_fails += is_dir_failed
-            if not is_dir_failed:
-                succeeded_dir_objs.append(dir_obj)
-
-    # insert the rest
-    total_fails += sum(processor(all_file_ids))
-    if total_fails:
-        if src_index:
-            src_index.clear()
-        raise FileTransferError(total_fails)
-
-    # index successfully pushed dirs
-    if dest_index:
-        for dir_obj in succeeded_dir_objs:
-            file_hashes = {oid.value for _, _, oid in dir_obj}
-            logger.debug(
-                "Indexing pushed dir '%s' with '%s' nested files",
-                dir_obj.hash_info,
-                len(file_hashes),
-            )
-            assert dir_obj.hash_info and dir_obj.hash_info.value
-            dest_index.update([dir_obj.hash_info.value], file_hashes)
-
-
-def transfer(
-    src: "ObjectDB",
-    dest: "ObjectDB",
-    obj_ids: Iterable["HashInfo"],
-    jobs: Optional[int] = None,
-    verify: bool = False,
-    hardlink: bool = False,
-    **kwargs,
-) -> int:
-    """Transfer (copy) the specified objects from one ODB to another.
-
-    Returns the number of successfully transferred objects
-    """
-    from .status import compare_status
-
-    logger.debug(
-        "Preparing to transfer data from '%s' to '%s'",
-        src.fs_path,
-        dest.fs_path,
-    )
-    if src == dest:
-        return 0
-
-    status = compare_status(
-        src, dest, obj_ids, check_deleted=False, jobs=jobs, **kwargs
-    )
-    if not status.new:
-        return 0
-
-    def func(hash_info: "HashInfo") -> None:
-        obj = src.get(hash_info)
-        return dest.add(
-            obj.fs_path,
-            obj.fs,
-            obj.hash_info,
-            verify=verify,
-            hardlink=hardlink,
-        )
-
-    total = len(status.new)
-    jobs = jobs or dest.fs.jobs
-    with Tqdm(total=total, unit="file", desc="Transferring") as pbar:
-        with ThreadPoolExecutor(max_workers=jobs) as executor:
-            wrapped_func = pbar.wrap_fn(_log_exceptions(func))
-            processor = partial(executor.imap_unordered, wrapped_func)
-            _do_transfer(
-                src, dest, status.new, status.missing, processor, **kwargs
-            )
-    return total

dvc/data/tree.py

@@ -1,293 +0,0 @@
-import json
-import logging
-import posixpath
-from typing import TYPE_CHECKING, Dict, Iterable, Optional, Tuple
-
-from funcy import cached_property
-
-from dvc.objects.errors import ObjectFormatError
-from dvc.objects.file import HashFile
-
-from .stage import get_file_hash
-
-if TYPE_CHECKING:
-    from dvc.hash_info import HashInfo
-    from dvc.objects.db import ObjectDB
-
-    from .meta import Meta
-
-logger = logging.getLogger(__name__)
-
-
-class TreeError(Exception):
-    pass
-
-
-def _try_load(
-    odbs: Iterable["ObjectDB"],
-    hash_info: "HashInfo",
-) -> Optional["HashFile"]:
-    for odb in odbs:
-        if not odb:
-            continue
-
-        try:
-            return Tree.load(odb, hash_info)
-        except (FileNotFoundError, ObjectFormatError):
-            pass
-
-    return None
-
-
-class Tree(HashFile):
-    PARAM_RELPATH = "relpath"
-
-    def __init__(self, *args, **kwargs):
-        super().__init__(*args, **kwargs)
-        self._dict: Dict[Tuple[str], Tuple["Meta", "HashInfo"]] = {}
-
-    @cached_property
-    def _trie(self):
-        from pygtrie import Trie
-
-        return Trie(self._dict)
-
-    def add(self, key: Tuple[str], meta: "Meta", oid: "HashInfo"):
-        self.__dict__.pop("trie", None)
-        self._dict[key] = (meta, oid)
-
-    def digest(self, hash_info: Optional["HashInfo"] = None):
-        from dvc.fs.memory import MemoryFileSystem
-        from dvc.utils import tmp_fname
-
-        memfs = MemoryFileSystem()
-        fs_path = "memory://{}".format(tmp_fname(""))
-        memfs.pipe_file(fs_path, self.as_bytes())
-        self.fs = memfs
-        self.fs_path = fs_path
-        if hash_info:
-            self.hash_info = hash_info
-        else:
-            _, self.hash_info = get_file_hash(fs_path, memfs, "md5")
-            assert self.hash_info.value
-            self.hash_info.value += ".dir"
-
-    def _load(self, key, meta, hash_info):
-        if hash_info and hash_info.isdir and not meta.obj:
-            meta.obj = _try_load([meta.odb, meta.remote], hash_info)
-            if meta.obj:
-                for ikey, value in meta.obj.iteritems():
-                    self._trie[key + ikey] = value
-                    self._dict[key + ikey] = value
-
-    def iteritems(self, prefix=None):
-        kwargs = {}
-        if prefix:
-            kwargs = {"prefix": prefix}
-            item = self._trie.longest_prefix(prefix)
-            if item:
-                key, (meta, hash_info) = item
-                self._load(key, meta, hash_info)
-
-        for key, (meta, hash_info) in self._trie.iteritems(**kwargs):
-            self._load(key, meta, hash_info)
-            yield key, (meta, hash_info)
-
-    def shortest_prefix(self, *args, **kwargs):
-        return self._trie.shortest_prefix(*args, **kwargs)
-
-    def __len__(self):
-        return len(self._dict)
-
-    def __iter__(self):
-        yield from (
-            (key, value[0], value[1]) for key, value in self._dict.items()
-        )
-
-    def as_dict(self):
-        return self._dict.copy()
-
-    def as_list(self):
-        from operator import itemgetter
-
-        # Sorting the list by path to ensure reproducibility
-        return sorted(
-            (
-                {
-                    # NOTE: not using hash_info.to_dict() because we don't want
-                    # size/nfiles fields at this point.
-                    oid.name: oid.value,
-                    self.PARAM_RELPATH: posixpath.sep.join(parts),
-                }
-                for parts, _, oid in self  # noqa: B301
-            ),
-            key=itemgetter(self.PARAM_RELPATH),
-        )
-
-    def as_bytes(self):
-        return json.dumps(self.as_list(), sort_keys=True).encode("utf-8")
-
-    @classmethod
-    def from_list(cls, lst):
-        from dvc.hash_info import HashInfo
-
-        tree = cls(None, None, None)
-        for _entry in lst:
-            entry = _entry.copy()
-            relpath = entry.pop(cls.PARAM_RELPATH)
-            parts = tuple(relpath.split(posixpath.sep))
-            hash_info = HashInfo.from_dict(entry)
-            tree.add(parts, None, hash_info)
-        return tree
-
-    @classmethod
-    def load(cls, odb, hash_info):
-        obj = odb.get(hash_info)
-
-        try:
-            with obj.fs.open(obj.fs_path, "r") as fobj:
-                raw = json.load(fobj)
-        except ValueError as exc:
-            raise ObjectFormatError(f"{obj} is corrupted") from exc
-
-        if not isinstance(raw, list):
-            logger.error(
-                "dir cache file format error '%s' [skipping the file]",
-                obj.fs_path,
-            )
-            raise ObjectFormatError(f"{obj} is corrupted")
-
-        tree = cls.from_list(raw)
-        tree.fs_path = obj.fs_path
-        tree.fs = obj.fs
-        tree.hash_info = hash_info
-
-        return tree
-
-    def filter(self, prefix: Tuple[str]) -> Optional["Tree"]:
-        """Return a filtered copy of this tree that only contains entries
-        inside prefix.
-
-        The returned tree will contain the original tree's hash_info and
-        fs_path.
-
-        Returns an empty tree if no object exists at the specified prefix.
-        """
-        tree = Tree(self.fs_path, self.fs, self.hash_info)
-        try:
-            for key, (meta, oid) in self._trie.items(prefix):
-                tree.add(key, meta, oid)
-        except KeyError:
-            pass
-        return tree
-
-    def get(self, odb, prefix: Tuple[str]) -> Optional[HashFile]:
-        """Return object at the specified prefix in this tree.
-
-        Returns None if no object exists at the specified prefix.
-        """
-        _, oid = self._dict.get(prefix) or (None, None)
-        if oid:
-            return odb.get(oid)
-
-        tree = Tree(None, None, None)
-        depth = len(prefix)
-        try:
-            for key, (meta, entry_oid) in self._trie.items(prefix):
-                tree.add(key[depth:], meta, entry_oid)
-        except KeyError:
-            return None
-        tree.digest()
-        return tree
-
-    def ls(self, prefix=None):
-        kwargs = {}
-        if prefix:
-            kwargs["prefix"] = prefix
-
-        meta, hash_info = self._trie.get(prefix, (None, None))
-        if hash_info and hash_info.isdir and meta and not meta.obj:
-            raise TreeError
-
-        ret = []
-
-        def node_factory(_, key, children, *args):
-            if key == prefix:
-                list(children)
-            else:
-                ret.append(key[-1])
-
-        self._trie.traverse(node_factory, **kwargs)
-
-        return ret
-
-
-def du(odb, tree):
-    try:
-        return sum(
-            odb.fs.size(odb.hash_to_path(oid.value)) for _, _, oid in tree
-        )
-    except FileNotFoundError:
-        return None
-
-
-def _diff(ancestor, other, allow_removed=False):
-    from dictdiffer import diff
-
-    from dvc.exceptions import MergeError
-
-    allowed = ["add"]
-    if allow_removed:
-        allowed.append("remove")
-
-    result = list(diff(ancestor, other))
-    for typ, _, _ in result:
-        if typ not in allowed:
-            raise MergeError(
-                "unable to auto-merge directories with diff that contains "
-                f"'{typ}'ed files"
-            )
-    return result
-
-
-def _merge(ancestor, our, their):
-    import copy
-
-    from dictdiffer import patch
-
-    our_diff = _diff(ancestor, our)
-    if not our_diff:
-        return copy.deepcopy(their)
-
-    their_diff = _diff(ancestor, their)
-    if not their_diff:
-        return copy.deepcopy(our)
-
-    # make sure there are no conflicting files
-    _diff(our, their, allow_removed=True)
-
-    return patch(our_diff + their_diff, ancestor)
-
-
-def merge(odb, ancestor_info, our_info, their_info):
-    from . import load
-
-    assert our_info
-    assert their_info
-
-    if ancestor_info:
-        ancestor = load(odb, ancestor_info)
-    else:
-        ancestor = Tree(None, None, None)
-
-    our = load(odb, our_info)
-    their = load(odb, their_info)
-
-    merged_dict = _merge(ancestor.as_dict(), our.as_dict(), their.as_dict())
-
-    merged = Tree(None, None, None)
-    for key, (meta, oid) in merged_dict.items():
-        merged.add(key, meta, oid)
-    merged.digest()
-
-    return merged

dvc/data_cloud.py

@@ -3,11 +3,12 @@
 import logging
 from typing import TYPE_CHECKING, Iterable, Optional
 
-from dvc.data.db import get_index
+from dvc_data.db import get_index
 
 if TYPE_CHECKING:
-    from dvc.hash_info import HashInfo
-    from dvc.objects.db import ObjectDB
+    from dvc_data.hashfile.db import HashFileDB
+    from dvc_data.hashfile.hash_info import HashInfo
+    from dvc_data.status import CompareStatusResult
 
 logger = logging.getLogger(__name__)
 
@@ -30,7 +31,7 @@ class DataCloud:
         self,
         name: Optional[str] = None,
         command: str = "<command>",
-    ) -> "ObjectDB":
+    ) -> "HashFileDB":
         from dvc.config import NoRemoteError
 
         if not name:
@@ -55,19 +56,45 @@ class DataCloud:
         raise NoRemoteError(error_msg)
 
     def _init_odb(self, name):
-        from dvc.data.db import get_odb
         from dvc.fs import get_cloud_fs
+        from dvc_data.db import get_odb
 
         cls, config, fs_path = get_cloud_fs(self.repo, name=name)
         config["tmp_dir"] = self.repo.index_db_dir
         return get_odb(cls(**config), fs_path, **config)
 
+    def _log_missing(self, status: "CompareStatusResult"):
+        if status.missing:
+            missing_desc = "\n".join(
+                f"name: {hash_info.obj_name}, {hash_info}"
+                for hash_info in status.missing
+            )
+            logger.warning(
+                "Some of the cache files do not exist neither locally "
+                f"nor on remote. Missing cache files:\n{missing_desc}"
+            )
+
+    def transfer(
+        self,
+        src_odb: "HashFileDB",
+        dest_odb: "HashFileDB",
+        objs: Iterable["HashInfo"],
+        **kwargs,
+    ):
+        from dvc.exceptions import FileTransferError
+        from dvc_data.transfer import TransferError, transfer
+
+        try:
+            return transfer(src_odb, dest_odb, objs, **kwargs)
+        except TransferError as exc:
+            raise FileTransferError(exc.fails) from exc
+
     def push(
         self,
         objs: Iterable["HashInfo"],
         jobs: Optional[int] = None,
         remote: Optional[str] = None,
-        odb: Optional["ObjectDB"] = None,
+        odb: Optional["HashFileDB"] = None,
     ):
         """Push data items in a cloud-agnostic way.
 
@@ -78,17 +105,15 @@ class DataCloud:
                 By default remote from core.remote config option is used.
             odb: optional ODB to push to. Overrides remote.
         """
-        from dvc.data.transfer import transfer
-
-        if not odb:
-            odb = self.get_remote_odb(remote, "push")
-        return transfer(
+        odb = odb or self.get_remote_odb(remote, "push")
+        return self.transfer(
             self.repo.odb.local,
             odb,
             objs,
             jobs=jobs,
             dest_index=get_index(odb),
             cache_odb=self.repo.odb.local,
+            validate_status=self._log_missing,
         )
 
     def pull(
@@ -96,7 +121,7 @@ class DataCloud:
         objs: Iterable["HashInfo"],
         jobs: Optional[int] = None,
         remote: Optional[str] = None,
-        odb: Optional["ObjectDB"] = None,
+        odb: Optional["HashFileDB"] = None,
     ):
         """Pull data items in a cloud-agnostic way.
 
@@ -107,11 +132,8 @@ class DataCloud:
                 By default remote from core.remote config option is used.
             odb: optional ODB to pull from. Overrides remote.
         """
-        from dvc.data.transfer import transfer
-
-        if not odb:
-            odb = self.get_remote_odb(remote, "pull")
-        return transfer(
+        odb = odb or self.get_remote_odb(remote, "pull")
+        return self.transfer(
             odb,
             self.repo.odb.local,
             objs,
@@ -119,6 +141,7 @@ class DataCloud:
             src_index=get_index(odb),
             cache_odb=self.repo.odb.local,
             verify=odb.verify,
+            validate_status=self._log_missing,
         )
 
     def status(
@@ -126,7 +149,7 @@ class DataCloud:
         objs: Iterable["HashInfo"],
         jobs: Optional[int] = None,
         remote: Optional[str] = None,
-        odb: Optional["ObjectDB"] = None,
+        odb: Optional["HashFileDB"] = None,
         log_missing: bool = True,
     ):
         """Check status of data items in a cloud-agnostic way.
@@ -141,7 +164,7 @@ class DataCloud:
             log_missing: log warning messages if file doesn't exist
                 neither in cache, neither in cloud.
         """
-        from dvc.data.status import compare_status
+        from dvc_data.status import compare_status
 
         if not odb:
             odb = self.get_remote_odb(remote, "status")
@@ -157,5 +180,5 @@ class DataCloud:
 
     def get_url_for(self, remote, checksum):
         odb = self.get_remote_odb(remote)
-        path = odb.hash_to_path(checksum)
+        path = odb.oid_to_path(checksum)
         return odb.fs.unstrip_protocol(path)

dvc/dependency/param.py

@@ -8,8 +8,8 @@ import dpath.util
 from voluptuous import Any
 
 from dvc.exceptions import DvcException
-from dvc.hash_info import HashInfo
-from dvc.utils.serialize import LOADERS, ParseError
+from dvc.utils.serialize import ParseError, load_path
+from dvc_data.hashfile.hash_info import HashInfo
 
 from .base import Dependency
 
@@ -139,12 +139,9 @@ class ParamsDependency(Dependency):
             )
 
     def read_file(self):
-        _, ext = os.path.splitext(self.fs_path)
-        loader = LOADERS[ext]
-
         self.validate_filepath()
         try:
-            return loader(self.fs_path, fs=self.repo.fs)
+            return load_path(self.fs_path, self.repo.fs)
         except ParseError as exc:
             raise BadParamFileError(
                 f"Unable to read parameters from '{self}'"

dvc/dependency/repo.py

@@ -5,12 +5,14 @@ from typing import TYPE_CHECKING, Dict, Optional, Set, Tuple
 
 from voluptuous import Required
 
+from dvc.prompt import confirm
+
 from .base import Dependency
 
 if TYPE_CHECKING:
-    from dvc.hash_info import HashInfo
-    from dvc.objects.db import ObjectDB
-    from dvc.objects.file import HashFile
+    from dvc_data.hashfile.file import HashFile
+    from dvc_data.hashfile.hash_info import HashInfo
+    from dvc_objects.db import ObjectDB
 
 
 class RepoDependency(Dependency):
@@ -29,7 +31,7 @@ class RepoDependency(Dependency):
 
     def __init__(self, def_repo, stage, *args, **kwargs):
         self.def_repo = def_repo
-        self._staged_objs: Dict[str, "HashFile"] = {}
+        self._objs: Dict[str, "HashFile"] = {}
         super().__init__(stage, *args, **kwargs)
 
     def _parse_path(self, fs, fs_path):
@@ -61,7 +63,7 @@ class RepoDependency(Dependency):
         return {self.PARAM_PATH: self.def_path, self.PARAM_REPO: self.def_repo}
 
     def download(self, to, jobs=None):
-        from dvc.data.checkout import checkout
+        from dvc_data.checkout import checkout
 
         for odb, objs in self.get_used_objs().items():
             self.repo.cloud.pull(objs, jobs=jobs, odb=odb)
@@ -72,8 +74,9 @@ class RepoDependency(Dependency):
             to.fs,
             obj,
             self.repo.odb.local,
-            dvcignore=None,
+            ignore=None,
             state=self.repo.state,
+            prompt=confirm,
         )
 
     def update(self, rev=None):
@@ -98,16 +101,14 @@ class RepoDependency(Dependency):
         self, obj_only=False, **kwargs
     ) -> Tuple[Dict[Optional["ObjectDB"], Set["HashInfo"]], "HashFile"]:
         from dvc.config import NoRemoteError
-        from dvc.data.stage import stage
-        from dvc.data.tree import Tree, TreeError
         from dvc.exceptions import NoOutputOrStageError, PathMissingError
         from dvc.utils import as_posix
+        from dvc_data.build import build
+        from dvc_data.objects.tree import Tree, TreeError
 
         local_odb = self.repo.odb.local
         locked = kwargs.pop("locked", True)
-        with self._make_repo(
-            locked=locked, cache_dir=local_odb.cache_dir
-        ) as repo:
+        with self._make_repo(locked=locked, cache_dir=local_odb.path) as repo:
             used_obj_ids = defaultdict(set)
             rev = repo.get_rev()
             if locked and self.def_repo.get(self.PARAM_REV_LOCK) is None:
@@ -130,45 +131,45 @@ class RepoDependency(Dependency):
                     pass
 
             try:
-                staging, _, staged_obj = stage(
+                object_store, _, obj = build(
                     local_odb,
                     as_posix(self.def_path),
-                    repo.repo_fs,
+                    repo.dvcfs,
                     local_odb.fs.PARAM_CHECKSUM,
                 )
             except (FileNotFoundError, TreeError) as exc:
                 raise PathMissingError(
                     self.def_path, self.def_repo[self.PARAM_URL]
                 ) from exc
-            staging = copy(staging)
-            staging.read_only = True
+            object_store = copy(object_store)
+            object_store.read_only = True
 
-            self._staged_objs[rev] = staged_obj
-            used_obj_ids[staging].add(staged_obj.hash_info)
-            if isinstance(staged_obj, Tree):
-                used_obj_ids[staging].update(oid for _, _, oid in staged_obj)
-            return used_obj_ids, staged_obj
+            self._objs[rev] = obj
+            used_obj_ids[object_store].add(obj.hash_info)
+            if isinstance(obj, Tree):
+                used_obj_ids[object_store].update(oid for _, _, oid in obj)
+            return used_obj_ids, obj
 
     def _check_circular_import(self, odb, obj_ids):
-        from dvc.data.db.reference import ReferenceObjectDB
-        from dvc.data.tree import Tree
         from dvc.exceptions import CircularImportError
-        from dvc.fs.repo import RepoFileSystem
+        from dvc.fs.dvc import DvcFileSystem
+        from dvc_data.db.reference import ReferenceHashFileDB
+        from dvc_data.objects.tree import Tree
 
-        if not isinstance(odb, ReferenceObjectDB):
+        if not isinstance(odb, ReferenceHashFileDB):
             return
 
         def iter_objs():
             for hash_info in obj_ids:
                 if hash_info.isdir:
                     tree = Tree.load(odb, hash_info)
-                    yield from (odb.get(oid) for _, _, oid in tree)
+                    yield from (odb.get(hi.value) for _, _, hi in tree)
                 else:
-                    yield odb.get(hash_info)
+                    yield odb.get(hash_info.value)
 
         checked_urls = set()
         for obj in iter_objs():
-            if not isinstance(obj.fs, RepoFileSystem):
+            if not isinstance(obj.fs, DvcFileSystem):
                 continue
             if (
                 obj.fs.repo_url in checked_urls
@@ -187,8 +188,8 @@ class RepoDependency(Dependency):
     def get_obj(self, filter_info=None, **kwargs):
         locked = kwargs.get("locked", True)
         rev = self._get_rev(locked=locked)
-        if rev in self._staged_objs:
-            return self._staged_objs[rev]
+        if rev in self._objs:
+            return self._objs[rev]
         _, obj = self._get_used_and_obj(
             obj_only=True, filter_info=filter_info, **kwargs
         )

dvc/dvcfile.py

@@ -59,7 +59,7 @@ def is_lock_file(path):
 
 
 def is_git_ignored(repo, path):
-    from dvc.fs.local import LocalFileSystem
+    from dvc.fs import LocalFileSystem
     from dvc.scm import NoSCMError
 
     try:
@@ -197,7 +197,7 @@ class SingleStageFile(FileMixin):
         assert not isinstance(stage, PipelineStage)
         if self.verify:
             check_dvcfile_path(self.repo, self.path)
-        logger.debug(f"Saving information to '{relpath(self.path)}'.")
+        logger.debug("Saving information to '%s'.", relpath(self.path))
         dump_yaml(self.path, serialize.to_single_stage_file(stage))
         self.repo.scm_context.track_file(self.relpath)
 

dvc/exceptions.py

@@ -202,14 +202,6 @@ class FileMissingError(DvcException):
         )
 
 
-class DvcIgnoreInCollectedDirError(DvcException):
-    def __init__(self, ignore_dirname):
-        super().__init__(
-            ".dvcignore file should not be in collected dir path: "
-            "'{}'".format(ignore_dirname)
-        )
-
-
 class FileTransferError(DvcException):
     _METHOD = "transfer"
 

dvc/external_repo.py

@@ -33,7 +33,7 @@ def external_repo(
     from scmrepo.git import Git
 
     from dvc.config import NoRemoteError
-    from dvc.fs.git import GitFileSystem
+    from dvc.fs import GitFileSystem
 
     logger.debug("Creating external repo %s@%s", url, rev)
     path = _cached_clone(url, rev, for_write=for_write)
@@ -50,19 +50,21 @@ def external_repo(
     config.update(cache_config)
 
     if for_write:
+        scm = None
         root_dir = path
         fs = None
     else:
-        root_dir = os.path.realpath(path)
-        scm = Git(root_dir)
+        scm = Git(path)
         fs = GitFileSystem(scm=scm, rev=rev)
+        root_dir = "/"
 
     repo_kwargs = dict(
         root_dir=root_dir,
         url=url,
         fs=fs,
         config=config,
-        repo_factory=erepo_factory(url, cache_config),
+        repo_factory=erepo_factory(url, root_dir, cache_config),
+        scm=scm,
         **kwargs,
     )
 
@@ -92,14 +94,16 @@ def external_repo(
             _remove(path)
 
 
-def erepo_factory(url, cache_config, *args, **kwargs):
-    def make_repo(path, **_kwargs):
+def erepo_factory(url, root_dir, cache_config):
+    from dvc.fs import localfs
+
+    def make_repo(path, fs=None, **_kwargs):
         _config = cache_config.copy()
         if os.path.isdir(url):
-            rel = os.path.relpath(path, _kwargs["fs"].fs_args["scm"].root_dir)
-            repo_path = os.path.join(url, rel)
+            fs = fs or localfs
+            repo_path = os.path.join(url, *fs.path.relparts(path, root_dir))
             _config.update(_get_remote_config(repo_path))
-        return Repo(path, config=_config, **_kwargs)
+        return Repo(path, fs=fs, config=_config, **_kwargs)
 
     return make_repo
 
@@ -141,7 +145,7 @@ def _get_remote_config(url):
             name = "auto-generated-upstream"
             return {
                 "core": {"remote": name},
-                "remote": {name: {"url": repo.odb.local.cache_dir}},
+                "remote": {name: {"url": repo.odb.local.path}},
             }
 
         # Use original remote to make sure that we are using correct url,

dvc/fs/__init__.py

@@ -1,41 +1,47 @@
 from urllib.parse import urlparse
 
-from ..scheme import Schemes
-from . import utils  # noqa: F401
-from .azure import AzureFileSystem
-from .gdrive import GDriveFileSystem
-from .gs import GSFileSystem
-from .hdfs import HDFSFileSystem
-from .http import HTTPFileSystem
-from .https import HTTPSFileSystem
-from .local import LocalFileSystem
-from .oss import OSSFileSystem
-from .s3 import S3FileSystem
-from .ssh import SSHFileSystem
-from .webdav import WebDAVFileSystem, WebDAVSFileSystem
-from .webhdfs import WebHDFSFileSystem
+from dvc_http import HTTPFileSystem, HTTPSFileSystem  # noqa: F401
 
-FS_MAP = {
-    Schemes.AZURE: AzureFileSystem,
-    Schemes.GDRIVE: GDriveFileSystem,
-    Schemes.GS: GSFileSystem,
-    Schemes.HDFS: HDFSFileSystem,
-    Schemes.WEBHDFS: WebHDFSFileSystem,
-    Schemes.HTTP: HTTPFileSystem,
-    Schemes.HTTPS: HTTPSFileSystem,
-    Schemes.S3: S3FileSystem,
-    Schemes.SSH: SSHFileSystem,
-    Schemes.OSS: OSSFileSystem,
-    Schemes.WEBDAV: WebDAVFileSystem,
-    Schemes.WEBDAVS: WebDAVSFileSystem,
-    # NOTE: LocalFileSystem is the default
-}
+# pylint: disable=unused-import
+from dvc_objects.fs import utils  # noqa: F401
+from dvc_objects.fs import (  # noqa: F401
+    LocalFileSystem,
+    MemoryFileSystem,
+    Schemes,
+    generic,
+    get_fs_cls,
+    known_implementations,
+    localfs,
+    registry,
+    system,
+)
+from dvc_objects.fs.base import AnyFSPath, FileSystem  # noqa: F401
+from dvc_objects.fs.errors import (  # noqa: F401
+    AuthError,
+    ConfigError,
+    RemoteMissingDepsError,
+)
+from dvc_objects.fs.path import Path  # noqa: F401
+
+from .data import DataFileSystem  # noqa: F401
+from .dvc import DvcFileSystem  # noqa: F401
+from .git import GitFileSystem  # noqa: F401
+
+known_implementations.update(
+    {
+        "dvc": {
+            "class": "dvc.fs.dvc.DvcFileSystem",
+            "err": "dvc is supported, but requires 'dvc' to be installed",
+        },
+        "git": {
+            "class": "dvc.fs.git.GitFileSystem",
+            "err": "git is supported, but requires 'dvc' to be installed",
+        },
+    }
+)
 
 
-def get_fs_cls(remote_conf, scheme=None):
-    if not scheme:
-        scheme = urlparse(remote_conf["url"]).scheme
-    return FS_MAP.get(scheme, LocalFileSystem)
+# pylint: enable=unused-import
 
 
 def get_fs_config(repo, config, **kwargs):
@@ -86,7 +92,7 @@ def _resolve_remote_refs(repo, config, remote_conf):
 
 
 def get_cloud_fs(repo, **kwargs):
-    from dvc.config import ConfigError
+    from dvc.config import ConfigError as RepoConfigError
     from dvc.config_schema import SCHEMA, Invalid
 
     repo_config = repo.config if repo else {}
@@ -96,12 +102,7 @@ def get_cloud_fs(repo, **kwargs):
     try:
         remote_conf = SCHEMA["remote"][str](remote_conf)
     except Invalid as exc:
-        raise ConfigError(str(exc)) from None
-
-    if "jobs" not in remote_conf:
-        jobs = core_config.get("jobs")
-        if jobs:
-            remote_conf["jobs"] = jobs
+        raise RepoConfigError(str(exc)) from None
 
     if "checksum_jobs" not in remote_conf:
         checksum_jobs = core_config.get("checksum_jobs")
@@ -110,11 +111,8 @@ def get_cloud_fs(repo, **kwargs):
 
     cls = get_fs_cls(remote_conf)
 
-    if cls == GDriveFileSystem and repo:
-        remote_conf["gdrive_credentials_tmp_dir"] = repo.tmp_dir
-
     url = remote_conf.pop("url")
-    if issubclass(cls, WebDAVFileSystem):
+    if cls.protocol in ["webdav", "webdavs"]:
         # For WebDAVFileSystem, provided url is the base path itself, so it
         # should be treated as being a root path.
         fs_path = cls.root_marker

dvc/fs/_callback.py

@@ -1,190 +0,0 @@
-from functools import wraps
-from typing import IO, TYPE_CHECKING, Any, Optional, TypeVar, cast
-
-import fsspec
-
-if TYPE_CHECKING:
-    from typing import Callable
-
-    from typing_extensions import ParamSpec
-
-    from dvc.fs.base import FileSystem
-    from dvc.progress import Tqdm
-    from dvc.ui._rich_progress import RichTransferProgress
-
-    _P = ParamSpec("_P")
-    _R = TypeVar("_R")
-
-
-class FsspecCallback(fsspec.Callback):
-    """FsspecCallback usable as a context manager, and a few helper methods."""
-
-    def wrap_attr(self, fobj: IO, method: str = "read") -> IO:
-        from tqdm.utils import CallbackIOWrapper
-
-        wrapped = CallbackIOWrapper(self.relative_update, fobj, method)
-        return cast(IO, wrapped)
-
-    def wrap_fn(self, fn: "Callable[_P, _R]") -> "Callable[_P, _R]":
-        @wraps(fn)
-        def wrapped(*args: "_P.args", **kwargs: "_P.kwargs") -> "_R":
-            res = fn(*args, **kwargs)
-            self.relative_update()
-            return res
-
-        return wrapped
-
-    def wrap_and_branch(
-        self, fn: "Callable", fs: "FileSystem" = None
-    ) -> "Callable":
-        """
-        Wraps a function, and pass a new child callback to it.
-        When the function completes, we increment the parent callback by 1.
-        """
-        from .local import localfs
-
-        fs = fs or localfs
-        wrapped = self.wrap_fn(fn)
-
-        def make_callback(path1, path2):
-            # pylint: disable=assignment-from-none
-            child = self.branch(fs.path.name(path1), path2, {})
-            return self.as_callback(child)
-
-        @wraps(fn)
-        def func(path1, path2, **kwargs):
-            with make_callback(path1, path2) as callback:
-                return wrapped(path1, path2, callback=callback, **kwargs)
-
-        return func
-
-    def __enter__(self):
-        return self
-
-    def __exit__(self, *exc_args):
-        self.close()
-
-    def close(self):
-        """Handle here on exit."""
-
-    @classmethod
-    def as_callback(
-        cls, maybe_callback: Optional["FsspecCallback"] = None
-    ) -> "FsspecCallback":
-        if maybe_callback is None:
-            return DEFAULT_CALLBACK
-        return maybe_callback
-
-    @classmethod
-    def as_tqdm_callback(
-        cls,
-        callback: Optional["FsspecCallback"] = None,
-        **tqdm_kwargs: Any,
-    ) -> "FsspecCallback":
-        return callback or TqdmCallback(**tqdm_kwargs)
-
-    @classmethod
-    def as_rich_callback(
-        cls, callback: Optional["FsspecCallback"] = None, **rich_kwargs
-    ):
-        return callback or RichCallback(**rich_kwargs)
-
-
-class NoOpCallback(FsspecCallback, fsspec.callbacks.NoOpCallback):
-    pass
-
-
-class TqdmCallback(FsspecCallback):
-    def __init__(self, progress_bar: "Tqdm" = None, **tqdm_kwargs):
-        from dvc.progress import Tqdm
-
-        self.progress_bar = progress_bar or Tqdm(**tqdm_kwargs)
-        super().__init__()
-
-    def __enter__(self):
-        self.progress_bar.__enter__()
-        return self
-
-    def close(self):
-        self.progress_bar.close()
-
-    def set_size(self, size):
-        if size is not None:
-            self.progress_bar.total = size
-            self.progress_bar.refresh()
-            super().set_size(size)
-
-    def relative_update(self, inc=1):
-        self.progress_bar.update(inc)
-        super().relative_update(inc)
-
-    def absolute_update(self, value):
-        self.progress_bar.update_to(value)
-        super().absolute_update(value)
-
-    def branch(self, path_1: str, path_2: str, kwargs):
-        return TqdmCallback(bytes=True, total=-1, desc=path_1, **kwargs)
-
-
-class RichCallback(FsspecCallback):
-    def __init__(
-        self,
-        progress: "RichTransferProgress" = None,
-        desc: str = None,
-        total: int = None,
-        bytes: bool = False,  # pylint: disable=redefined-builtin
-        unit: str = None,
-        disable: bool = False,
-    ) -> None:
-        from dvc.ui import ui
-        from dvc.ui._rich_progress import RichTransferProgress
-
-        self.progress = progress or RichTransferProgress(
-            transient=True,
-            disable=disable,
-            console=ui.error_console,
-        )
-        self.visible = not disable
-        self._newly_created = progress is None
-        total = 0 if total is None or total < 0 else total
-        self.task = self.progress.add_task(
-            description=desc or "",
-            total=total,
-            bytes=bytes,
-            visible=False,
-            unit=unit,
-            progress_type=None if bytes else "summary",
-        )
-        super().__init__()
-
-    def __enter__(self):
-        if self._newly_created:
-            self.progress.__enter__()
-        return self
-
-    def close(self):
-        if self._newly_created:
-            self.progress.stop()
-        try:
-            self.progress.remove_task(self.task)
-        except KeyError:
-            pass
-
-    def set_size(self, size: int = None) -> None:
-        if size is not None:
-            self.progress.update(self.task, total=size, visible=self.visible)
-            super().set_size(size)
-
-    def relative_update(self, inc: int = 1) -> None:
-        self.progress.update(self.task, advance=inc)
-        super().relative_update(inc)
-
-    def absolute_update(self, value: int) -> None:
-        self.progress.update(self.task, completed=value)
-        super().absolute_update(value)
-
-    def branch(self, path_1, path_2, kwargs):
-        return RichCallback(self.progress, desc=path_1, bytes=True, total=-1)
-
-
-DEFAULT_CALLBACK = NoOpCallback()

dvc/fs/azure.py

@@ -1,160 +0,0 @@
-import logging
-import os
-import threading
-
-from fsspec.asyn import fsspec_loop
-from fsspec.utils import infer_storage_options
-from funcy import cached_property, memoize, wrap_prop
-
-from dvc.exceptions import DvcException
-from dvc.scheme import Schemes
-from dvc.utils import format_link
-
-from .base import ObjectFileSystem
-
-logger = logging.getLogger(__name__)
-_DEFAULT_CREDS_STEPS = (
-    "https://azuresdkdocs.blob.core.windows.net/$web/python/"
-    "azure-identity/1.4.0/azure.identity.html#azure.identity"
-    ".DefaultAzureCredential"
-)
-
-
-class AzureAuthError(DvcException):
-    pass
-
-
-@memoize
-def _az_config():
-    # NOTE: ideally we would've used get_default_cli().config from
-    # azure.cli.core, but azure-cli-core has a lot of conflicts with other
-    # dependencies. So instead we are just use knack directly
-    from knack.config import CLIConfig
-
-    config_dir = os.getenv(
-        "AZURE_CONFIG_DIR", os.path.expanduser(os.path.join("~", ".azure"))
-    )
-    return CLIConfig(config_dir=config_dir, config_env_var_prefix="AZURE")
-
-
-# pylint:disable=abstract-method
-class AzureFileSystem(ObjectFileSystem):
-    scheme = Schemes.AZURE
-    PARAM_CHECKSUM = "etag"
-    REQUIRES = {
-        "adlfs": "adlfs",
-        "knack": "knack",
-        "azure-identity": "azure.identity",
-    }
-
-    @classmethod
-    def _strip_protocol(cls, path: str):
-        opts = infer_storage_options(path)
-        if opts.get("host"):
-            return "{host}{path}".format(**opts)
-
-        return _az_config().get("storage", "container_name", None)
-
-    def unstrip_protocol(self, path: str) -> str:
-        return "azure://" + path.lstrip("/")
-
-    @staticmethod
-    def _get_kwargs_from_urls(urlpath):
-        ops = infer_storage_options(urlpath)
-        if "host" in ops:
-            return {"bucket": ops["host"]}
-        return {}
-
-    def _prepare_credentials(self, **config):
-        from azure.identity.aio import DefaultAzureCredential
-
-        # Disable spam from failed cred types for DefaultAzureCredential
-        logging.getLogger("azure.identity.aio").setLevel(logging.ERROR)
-
-        login_info = {}
-        login_info["connection_string"] = config.get(
-            "connection_string",
-            _az_config().get("storage", "connection_string", None),
-        )
-        login_info["account_name"] = config.get(
-            "account_name", _az_config().get("storage", "account", None)
-        )
-        login_info["account_key"] = config.get(
-            "account_key", _az_config().get("storage", "key", None)
-        )
-        login_info["sas_token"] = config.get(
-            "sas_token", _az_config().get("storage", "sas_token", None)
-        )
-        login_info["tenant_id"] = config.get("tenant_id")
-        login_info["client_id"] = config.get("client_id")
-        login_info["client_secret"] = config.get("client_secret")
-
-        if not (login_info["account_name"] or login_info["connection_string"]):
-            raise AzureAuthError(
-                "Authentication to Azure Blob Storage requires either "
-                "account_name or connection_string.\nLearn more about "
-                "configuration settings at "
-                + format_link("https://man.dvc.org/remote/modify")
-            )
-
-        any_secondary = any(
-            value for key, value in login_info.items() if key != "account_name"
-        )
-        if (
-            login_info["account_name"]
-            and not any_secondary
-            and not config.get("allow_anonymous_login", False)
-        ):
-            with fsspec_loop():
-                login_info["credential"] = DefaultAzureCredential(
-                    exclude_interactive_browser_credential=False,
-                    exclude_environment_credential=config.get(
-                        "exclude_environment_credential", False
-                    ),
-                    exclude_visual_studio_code_credential=config.get(
-                        "exclude_visual_studio_code_credential", False
-                    ),
-                    exclude_shared_token_cache_credential=config.get(
-                        "exclude_shared_token_cache_credential", False
-                    ),
-                    exclude_managed_identity_credential=config.get(
-                        "exclude_managed_identity_credential", False
-                    ),
-                )
-
-        for login_method, required_keys in [  # noqa
-            ("connection string", ["connection_string"]),
-            (
-                "AD service principal",
-                ["tenant_id", "client_id", "client_secret"],
-            ),
-            ("account key", ["account_name", "account_key"]),
-            ("SAS token", ["account_name", "sas_token"]),
-            (
-                f"default credentials ({_DEFAULT_CREDS_STEPS})",
-                ["account_name", "credential"],
-            ),
-            ("anonymous login", ["account_name"]),
-        ]:
-            if all(login_info.get(key) is not None for key in required_keys):
-                break
-        else:
-            login_method = None
-
-        self.login_method = login_method
-        return login_info
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        from adlfs import AzureBlobFileSystem
-        from azure.core.exceptions import AzureError
-
-        try:
-            return AzureBlobFileSystem(**self.fs_args)
-        except (ValueError, AzureError) as e:
-            raise AzureAuthError(
-                f"Authentication to Azure Blob Storage via {self.login_method}"
-                " failed.\nLearn more about configuration settings at"
-                f" {format_link('https://man.dvc.org/remote/modify')}"
-            ) from e

dvc/fs/base.py

@@ -1,504 +0,0 @@
-import logging
-import os
-import shutil
-from multiprocessing import cpu_count
-from typing import (
-    IO,
-    TYPE_CHECKING,
-    Any,
-    ClassVar,
-    Dict,
-    Iterator,
-    List,
-    Optional,
-    Union,
-    cast,
-    overload,
-)
-
-from funcy import cached_property
-
-from dvc.exceptions import DvcException
-from dvc.utils.fs import as_atomic, makedirs
-from dvc.utils.threadpool import ThreadPoolExecutor
-
-from ._callback import DEFAULT_CALLBACK, FsspecCallback
-
-if TYPE_CHECKING:
-    from fsspec.spec import AbstractFileSystem
-    from typing_extensions import Literal
-
-
-logger = logging.getLogger(__name__)
-
-
-FSPath = str
-AnyFSPath = str
-
-# An info() entry, might evolve to a TypedDict
-# in the future (e.g for properly type 'size' etc).
-Entry = Dict[str, Any]
-
-
-class RemoteActionNotImplemented(DvcException):
-    def __init__(self, action, scheme):
-        m = f"{action} is not supported for {scheme} remotes"
-        super().__init__(m)
-
-
-class RemoteMissingDepsError(DvcException):
-    pass
-
-
-class FileSystem:
-    sep = "/"
-
-    scheme = "base"
-    REQUIRES: ClassVar[Dict[str, str]] = {}
-    _JOBS = 4 * cpu_count()
-
-    HASH_JOBS = max(1, min(4, cpu_count() // 2))
-    LIST_OBJECT_PAGE_SIZE = 1000
-    TRAVERSE_WEIGHT_MULTIPLIER = 5
-    TRAVERSE_PREFIX_LEN = 2
-    TRAVERSE_THRESHOLD_SIZE = 500000
-    CAN_TRAVERSE = True
-
-    PARAM_CHECKSUM: ClassVar[Optional[str]] = None
-
-    def __init__(self, **kwargs):
-        self._check_requires(**kwargs)
-
-        self.jobs = kwargs.get("jobs") or self._JOBS
-        self.hash_jobs = kwargs.get("checksum_jobs") or self.HASH_JOBS
-        self._config = kwargs
-        self.fs_args = {"skip_instance_cache": True}
-        self.fs_args.update(self._prepare_credentials(**kwargs))
-
-    @property
-    def config(self) -> Dict[str, Any]:
-        return self._config
-
-    @cached_property
-    def path(self):
-        from .path import Path
-
-        return Path(self.sep)
-
-    @classmethod
-    def _strip_protocol(cls, path: str) -> str:
-        return path
-
-    def unstrip_protocol(self, path: str) -> str:
-        return path
-
-    @cached_property
-    def fs(self) -> "AbstractFileSystem":
-        raise NotImplementedError
-
-    @staticmethod
-    def _get_kwargs_from_urls(urlpath: str) -> "Dict[str, Any]":
-        from fsspec.utils import infer_storage_options
-
-        options = infer_storage_options(urlpath)
-        options.pop("path", None)
-        options.pop("protocol", None)
-        return options
-
-    def _prepare_credentials(
-        self, **config: Dict[str, Any]  # pylint: disable=unused-argument
-    ) -> Dict[str, Any]:
-        """Prepare the arguments for authentication to the
-        host filesystem"""
-        return {}
-
-    @classmethod
-    def get_missing_deps(cls) -> List[str]:
-        import importlib
-
-        missing: List[str] = []
-        for package, module in cls.REQUIRES.items():
-            try:
-                importlib.import_module(module)
-            except ImportError:
-                missing.append(package)
-
-        return missing
-
-    def _check_requires(self, **kwargs):
-        from ..scheme import Schemes
-        from ..utils import format_link
-        from ..utils.pkg import PKG
-
-        missing = self.get_missing_deps()
-        if not missing:
-            return
-
-        url = kwargs.get("url", f"{self.scheme}://")
-
-        scheme = self.scheme
-        if scheme == Schemes.WEBDAVS:
-            scheme = Schemes.WEBDAV
-
-        by_pkg = {
-            "pip": f"pip install 'dvc[{scheme}]'",
-            "conda": f"conda install -c conda-forge dvc-{scheme}",
-        }
-
-        cmd = by_pkg.get(PKG)
-        if cmd:
-            link = format_link("https://dvc.org/doc/install")
-            hint = (
-                f"To install dvc with those dependencies, run:\n"
-                "\n"
-                f"\t{cmd}\n"
-                "\n"
-                f"See {link} for more info."
-            )
-        else:
-            link = format_link("https://github.com/iterative/dvc/issues")
-            hint = f"Please report this bug to {link}. Thank you!"
-
-        raise RemoteMissingDepsError(
-            f"URL '{url}' is supported but requires these missing "
-            f"dependencies: {missing}. {hint}"
-        )
-
-    def isdir(self, path: AnyFSPath) -> bool:
-        return self.fs.isdir(path)
-
-    def isfile(self, path: AnyFSPath) -> bool:
-        return self.fs.isfile(path)
-
-    def is_empty(self, path: AnyFSPath) -> bool:
-        entry = self.info(path)
-        if entry["type"] == "directory":
-            return not self.fs.ls(path)
-        return entry["size"] == 0
-
-    def open(
-        self,
-        path: AnyFSPath,
-        mode: str = "r",
-        encoding: Optional[str] = None,
-        **kwargs,
-    ) -> "IO":  # pylint: disable=arguments-differ
-        return self.fs.open(path, mode=mode, encoding=encoding, **kwargs)
-
-    def cat(
-        self,
-        path: Union[AnyFSPath, List[AnyFSPath]],
-        recursive: bool = False,
-        on_error: "Literal['raise', 'omit', 'return']" = "raise",
-        **kwargs: Any,
-    ) -> Union[bytes, Dict[AnyFSPath, bytes]]:
-        return self.fs.cat(
-            path, recursive=recursive, on_error=on_error, **kwargs
-        )
-
-    def cat_ranges(
-        self,
-        paths: List[AnyFSPath],
-        starts: List[int],
-        ends: List[int],
-        max_gap: int = None,
-        **kwargs,
-    ) -> List[bytes]:
-        return self.fs.cat_ranges(
-            paths, starts, ends, max_gap=max_gap, **kwargs
-        )
-
-    def cat_file(
-        self,
-        path: AnyFSPath,
-        start: int = None,
-        end: int = None,
-        **kwargs: Any,
-    ) -> bytes:
-        return self.fs.cat_file(path, start=start, end=end, **kwargs)
-
-    def pipe_file(self, path: AnyFSPath, value: bytes, **kwargs: Any) -> None:
-        return self.fs.pipe_file(path, value, **kwargs)
-
-    def pipe(
-        self,
-        path: Union[AnyFSPath, Dict[AnyFSPath, bytes]],
-        value: Optional[bytes] = None,
-        **kwargs: Any,
-    ) -> None:
-        return self.fs.pipe(path, value=value, **kwargs)
-
-    def touch(
-        self, path: AnyFSPath, truncate: bool = True, **kwargs: Any
-    ) -> None:
-        return self.fs.touch(path, truncate=truncate, **kwargs)
-
-    def checksum(self, path: AnyFSPath) -> str:
-        return self.fs.checksum(path)
-
-    def copy(self, from_info: AnyFSPath, to_info: AnyFSPath) -> None:
-        self.makedirs(self.path.parent(to_info))
-        self.fs.copy(from_info, to_info)
-
-    def exists(self, path: AnyFSPath) -> bool:
-        return self.fs.exists(path)
-
-    def lexists(self, path: AnyFSPath) -> bool:
-        return self.fs.lexists(path)
-
-    def symlink(self, from_info: AnyFSPath, to_info: AnyFSPath) -> None:
-        try:
-            return self.fs.symlink(from_info, to_info)
-        except AttributeError:
-            raise RemoteActionNotImplemented("symlink", self.scheme)
-
-    def hardlink(self, from_info: AnyFSPath, to_info: AnyFSPath) -> None:
-        try:
-            return self.fs.hardlink(from_info, to_info)
-        except AttributeError:
-            raise RemoteActionNotImplemented("hardlink", self.scheme)
-
-    def reflink(self, from_info: AnyFSPath, to_info: AnyFSPath) -> None:
-        try:
-            return self.fs.reflink(from_info, to_info)
-        except AttributeError:
-            raise RemoteActionNotImplemented("reflink", self.scheme)
-
-    def is_symlink(self, path: AnyFSPath) -> bool:
-        try:
-            return self.fs.is_symlink(path)
-        except AttributeError:
-            return False
-
-    def is_hardlink(self, path: AnyFSPath) -> bool:
-        try:
-            return self.fs.is_hardlink(path)
-        except AttributeError:
-            return False
-
-    def iscopy(self, path: AnyFSPath) -> bool:
-        return self.is_symlink(path) or self.is_hardlink(path)
-
-    @overload
-    def ls(
-        self, path: AnyFSPath, detail: "Literal[True]"
-    ) -> "Iterator[Entry]":
-        ...
-
-    @overload
-    def ls(self, path: AnyFSPath, detail: "Literal[False]") -> Iterator[str]:
-        ...
-
-    def ls(self, path, detail=False, **kwargs):
-        return self.fs.ls(path, detail=detail)
-
-    def find(
-        self,
-        path: AnyFSPath,
-        prefix: bool = False,  # pylint: disable=unused-argument
-    ) -> Iterator[str]:
-        yield from self.fs.find(path)
-
-    def move(self, from_info: AnyFSPath, to_info: AnyFSPath) -> None:
-        self.fs.move(from_info, to_info)
-
-    def remove(self, path: AnyFSPath) -> None:
-        self.fs.rm(path, recursive=True)
-
-    def info(self, path: AnyFSPath) -> "Entry":
-        return self.fs.info(path)
-
-    def makedirs(self, path: AnyFSPath, **kwargs: Any) -> None:
-        self.fs.makedirs(path, exist_ok=kwargs.pop("exist_ok", True))
-
-    def put_file(
-        self,
-        from_file: Union[AnyFSPath, IO],
-        to_info: AnyFSPath,
-        callback: FsspecCallback = DEFAULT_CALLBACK,
-        size: int = None,
-        **kwargs,
-    ) -> None:
-        if size:
-            callback.set_size(size)
-        if hasattr(from_file, "read"):
-            stream = callback.wrap_attr(cast("IO", from_file))
-            self.upload_fobj(stream, to_info, size=size)
-        else:
-            assert isinstance(from_file, str)
-            self.fs.put_file(
-                os.fspath(from_file), to_info, callback=callback, **kwargs
-            )
-        self.fs.invalidate_cache(self.path.parent(to_info))
-
-    def get_file(
-        self,
-        from_info: AnyFSPath,
-        to_info: AnyFSPath,
-        callback: FsspecCallback = DEFAULT_CALLBACK,
-        **kwargs,
-    ) -> None:
-        self.fs.get_file(from_info, to_info, callback=callback, **kwargs)
-
-    def upload_fobj(self, fobj: IO, to_info: AnyFSPath, **kwargs) -> None:
-        self.makedirs(self.path.parent(to_info))
-        with self.open(to_info, "wb") as fdest:
-            shutil.copyfileobj(
-                fobj,
-                fdest,
-                length=getattr(fdest, "blocksize", None),  # type: ignore
-            )
-
-    def walk(
-        self,
-        path: AnyFSPath,
-        topdown: bool = True,
-        **kwargs: Any,
-    ):
-        return self.fs.walk(path, topdown=topdown, **kwargs)
-
-    def size(self, path: AnyFSPath) -> Optional[int]:
-        return self.fs.size(path)
-
-    # pylint: enable=unused-argument
-
-    def upload(
-        self,
-        from_info: Union[AnyFSPath, IO],
-        to_info: AnyFSPath,
-        size: int = None,
-        callback: FsspecCallback = None,
-    ):
-        from .local import localfs
-
-        if not hasattr(from_info, "read"):
-            logger.debug("Uploading '%s' to '%s'", from_info, to_info)
-            desc = localfs.path.name(from_info)
-        else:
-            desc = self.path.name(to_info)
-
-        with FsspecCallback.as_tqdm_callback(
-            callback,
-            desc=desc,
-            bytes=True,
-            total=size or -1,
-        ) as cb:
-            return self.put_file(from_info, to_info, callback=cb, size=size)
-
-    def download(
-        self,
-        from_info: AnyFSPath,
-        to_info: AnyFSPath,
-        callback: "FsspecCallback" = None,
-        jobs: int = None,
-    ):
-        if not self.isdir(from_info):
-            return self.download_file(from_info, to_info, callback=callback)
-
-        from .local import localfs
-
-        pairs = {
-            info: localfs.path.join(
-                to_info, *self.path.relparts(info, from_info)
-            )
-            for info in self.find(from_info)
-        }
-        if not pairs:
-            makedirs(to_info, exist_ok=True)
-            return
-
-        with FsspecCallback.as_tqdm_callback(
-            callback,
-            total=-1,
-            desc=f"Downloading directory {self.path.name(from_info)}",
-            unit="files",
-        ) as cb:
-            cb.set_size(len(pairs))
-            download_files = cb.wrap_and_branch(self.download_file, fs=self)
-            max_workers = jobs or self.jobs
-            # NOTE: unlike pulling/fetching cache, where we need to
-            # download everything we can, not raising an error here might
-            # turn very ugly, as the user might think that he has
-            # downloaded a complete directory, while having a partial one,
-            # which might cause unexpected results in his pipeline.
-            with ThreadPoolExecutor(
-                max_workers=max_workers, cancel_on_error=True
-            ) as executor:
-                list(
-                    executor.imap_unordered(
-                        lambda args: download_files(*args), pairs.items()
-                    )
-                )
-
-    def download_file(
-        self,
-        from_info: AnyFSPath,
-        to_info: AnyFSPath,
-        callback: FsspecCallback = None,
-    ):
-        from .local import localfs
-
-        with FsspecCallback.as_tqdm_callback(
-            callback,
-            total=-1,
-            desc=self.path.name(from_info),
-            bytes=True,
-        ) as cb:
-            with as_atomic(localfs, to_info, create_parents=True) as tmp_file:
-                self.get_file(from_info, tmp_file, callback=cb)
-
-
-class ObjectFileSystem(FileSystem):  # pylint: disable=abstract-method
-    TRAVERSE_PREFIX_LEN = 3
-
-    def makedirs(self, path: AnyFSPath, **kwargs: Any) -> None:
-        # For object storages make this method a no-op. The original
-        # fs.makedirs() method will only check if the bucket exists
-        # and create if it doesn't though we don't want to support
-        # that behavior, and the check will cost some time so we'll
-        # simply ignore all mkdir()/makedirs() calls.
-        return None
-
-    def _isdir(self, path: AnyFSPath) -> bool:
-        # Directory in object storages are interpreted differently
-        # among different fsspec providers, so this logic is a temporary
-        # measure for us to adapt as of now. It checks whether it is a
-        # directory (as in a prefix with contents) or whether it is an empty
-        # file where it's name ends with a forward slash
-
-        entry = self.info(path)
-        return entry["type"] == "directory" or (
-            entry["size"] == 0
-            and entry["type"] == "file"
-            and entry["name"].endswith("/")
-        )
-
-    def isdir(self, path: AnyFSPath) -> bool:
-        try:
-            return self._isdir(path)
-        except FileNotFoundError:
-            return False
-
-    def isfile(self, path: AnyFSPath) -> bool:
-        try:
-            return not self._isdir(path)
-        except FileNotFoundError:
-            return False
-
-    def find(self, path: AnyFSPath, prefix: bool = False) -> Iterator[str]:
-        if prefix:
-            with_prefix = self.path.parent(path)
-            files = self.fs.find(with_prefix, prefix=self.path.parts(path)[-1])
-        else:
-            with_prefix = path
-            files = self.fs.find(path)
-
-        # When calling find() on a file, it returns the same file in a list.
-        # For object-based storages, the same behavior applies to empty
-        # directories since they are represented as files. This condition
-        # checks whether we should yield an empty list (if it is an empty
-        # directory) or just yield the file itself.
-        if len(files) == 1 and files[0] == with_prefix and self.isdir(path):
-            return None
-
-        yield from files

dvc/fs/callbacks.py

@@ -0,0 +1,79 @@
+# pylint: disable=unused-import
+from contextlib import ExitStack
+from typing import TYPE_CHECKING, Optional
+
+from funcy import cached_property
+
+from dvc_objects.fs.callbacks import (  # noqa: F401
+    DEFAULT_CALLBACK,
+    Callback,
+    TqdmCallback,
+)
+
+if TYPE_CHECKING:
+    from dvc.ui._rich_progress import RichTransferProgress
+
+
+class RichCallback(Callback):
+    def __init__(
+        self,
+        size: Optional[int] = None,
+        value: int = 0,
+        progress: "RichTransferProgress" = None,
+        desc: str = None,
+        bytes: bool = False,  # pylint: disable=redefined-builtin
+        unit: str = None,
+        disable: bool = False,
+    ) -> None:
+        self._progress = progress
+        self.disable = disable
+        self._task_kwargs = {
+            "description": desc or "",
+            "bytes": bytes,
+            "unit": unit,
+            "total": size or 0,
+            "visible": False,
+            "progress_type": None if bytes else "summary",
+        }
+        self._stack = ExitStack()
+        super().__init__(size=size, value=value)
+
+    @cached_property
+    def progress(self):
+        from dvc.ui import ui
+        from dvc.ui._rich_progress import RichTransferProgress
+
+        if self._progress is not None:
+            return self._progress
+
+        progress = RichTransferProgress(
+            transient=True,
+            disable=self.disable,
+            console=ui.error_console,
+        )
+        return self._stack.enter_context(progress)
+
+    @cached_property
+    def task(self):
+        return self.progress.add_task(**self._task_kwargs)
+
+    def __enter__(self):
+        return self
+
+    def close(self):
+        self.progress.clear_task(self.task)
+        self._stack.close()
+
+    def call(self, hook_name=None, **kwargs):
+        self.progress.update(
+            self.task,
+            completed=self.value,
+            total=self.size,
+            visible=not self.disable,
+        )
+
+    def branch(self, path_1, path_2, kwargs, child: Optional[Callback] = None):
+        child = child or RichCallback(
+            progress=self.progress, desc=path_1, bytes=True
+        )
+        return super().branch(path_1, path_2, kwargs, child=child)

dvc/fs/data.py

@@ -0,0 +1,34 @@
+import logging
+import os
+import threading
+
+from funcy import cached_property, wrap_prop
+
+from dvc.utils import as_posix
+from dvc_data.fs import DataFileSystem as _DataFileSystem
+from dvc_objects.fs.base import FileSystem
+
+logger = logging.getLogger(__name__)
+
+
+class DataFileSystem(FileSystem):
+    protocol = "local"
+
+    PARAM_CHECKSUM = "md5"
+
+    def _prepare_credentials(self, **config):
+        return config
+
+    @wrap_prop(threading.Lock())
+    @cached_property
+    def fs(self):
+        return _DataFileSystem(**self.fs_args)
+
+    def isdvc(self, path, **kwargs):
+        return self.fs.isdvc(path, **kwargs)
+
+    def from_os_path(self, path):
+        if os.path.isabs(path):
+            path = os.path.splitdrive(path)[1]
+
+        return as_posix(path)

dvc/fs/dvc.py

@@ -1,175 +1,447 @@
 import logging
+import ntpath
 import os
+import posixpath
 import threading
-import typing
+from contextlib import suppress
+from itertools import takewhile
+from typing import TYPE_CHECKING, Callable, Optional, Tuple, Type, Union
 
-from fsspec import AbstractFileSystem
-from funcy import cached_property, wrap_prop
+from fsspec.spec import AbstractFileSystem
+from funcy import cached_property, wrap_prop, wrap_with
 
-from ._callback import DEFAULT_CALLBACK
-from .base import FileSystem
+from dvc_objects.fs.base import FileSystem
+from dvc_objects.fs.callbacks import DEFAULT_CALLBACK
+from dvc_objects.fs.path import Path
 
-if typing.TYPE_CHECKING:
-    from dvc.types import AnyPath
+from .data import DataFileSystem
+
+if TYPE_CHECKING:
+    from dvc.repo import Repo
 
 logger = logging.getLogger(__name__)
 
+RepoFactory = Union[Callable[[str], "Repo"], Type["Repo"]]
+Key = Tuple[str, ...]
+
+
+def as_posix(path: str) -> str:
+    return path.replace(ntpath.sep, posixpath.sep)
+
+
+def _wrap_walk(dvc_fs, *args, **kwargs):
+    for root, dnames, fnames in dvc_fs.walk(*args, **kwargs):
+        yield dvc_fs.path.join(dvc_fs.repo.root_dir, root), dnames, fnames
+
+
+# NOT the same as dvc.dvcfile.is_dvc_file()!
+def _is_dvc_file(fname):
+    from dvc.dvcfile import is_valid_filename
+    from dvc.ignore import DvcIgnore
+
+    return is_valid_filename(fname) or fname == DvcIgnore.DVCIGNORE_FILE
+
+
+def _ls(fs, path):
+    dnames = []
+    fnames = []
+
+    for entry in fs.ls(path, detail=True):
+        name = fs.path.name(entry["name"])
+        if entry["type"] == "directory":
+            dnames.append(name)
+        else:
+            fnames.append(name)
+
+    return dnames, fnames
+
+
+def _merge_info(repo, fs_info, dvc_info):
+    from . import utils
+
+    ret = {"repo": repo}
+
+    if dvc_info:
+        ret["dvc_info"] = dvc_info
+        ret["type"] = dvc_info["type"]
+        ret["size"] = dvc_info["size"]
+        if not fs_info and "md5" in dvc_info:
+            ret["md5"] = dvc_info["md5"]
+
+    if fs_info:
+        ret["type"] = fs_info["type"]
+        ret["size"] = fs_info["size"]
+        isexec = False
+        if fs_info["type"] == "file":
+            isexec = utils.is_exec(fs_info["mode"])
+        ret["isexec"] = isexec
+
+    return ret
+
 
 class _DvcFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
-    """DVC repo fs.
+    """DVC + git-tracked files fs.
 
     Args:
-        repo: DVC repo.
+        repo: DVC or git repo.
+        subrepos: traverse to subrepos (by default, it ignores subrepos)
+        repo_factory: A function to initialize subrepo with, default is Repo.
+        kwargs: Additional keyword arguments passed to the `DataFileSystem()`.
     """
 
-    def __init__(self, **kwargs):
-        super().__init__(**kwargs)
-        self.repo = kwargs["repo"]
+    root_marker = "/"
+
+    PARAM_REPO_URL = "repo_url"
+    PARAM_REPO_ROOT = "repo_root"
+    PARAM_REV = "rev"
+    PARAM_CACHE_DIR = "cache_dir"
+    PARAM_CACHE_TYPES = "cache_types"
+    PARAM_SUBREPOS = "subrepos"
+
+    def __init__(
+        self,
+        repo: Optional["Repo"] = None,
+        subrepos=False,
+        repo_factory: RepoFactory = None,
+        **kwargs,
+    ):
+        super().__init__()
+
+        from pygtrie import Trie
+
+        if repo is None:
+            repo, repo_factory = self._repo_from_fs_config(
+                subrepos=subrepos, **kwargs
+            )
+
+        if not repo_factory:
+            from dvc.repo import Repo
+
+            self.repo_factory: RepoFactory = Repo
+        else:
+            self.repo_factory = repo_factory
+
+        def _getcwd():
+            relparts = ()
+            if repo.fs.path.isin(repo.fs.path.getcwd(), repo.root_dir):
+                relparts = repo.fs.path.relparts(
+                    repo.fs.path.getcwd(), repo.root_dir
+                )
+            return self.root_marker + self.sep.join(relparts)
+
+        self.path = Path(self.sep, getcwd=_getcwd)
+        self.repo = repo
+        self.hash_jobs = repo.fs.hash_jobs
+        self._traverse_subrepos = subrepos
+
+        self._subrepos_trie = Trie()
+        """Keeps track of each and every path with the corresponding repo."""
+
+        key = self._get_key(self.repo.root_dir)
+        self._subrepos_trie[key] = repo
+
+        self._datafss = {}
+        """Keep a datafs instance of each repo."""
+
+        if hasattr(repo, "dvc_dir"):
+            self._datafss[key] = DataFileSystem(index=repo.index.data["repo"])
+
+    def _get_key(self, path) -> Key:
+        parts = self.repo.fs.path.relparts(path, self.repo.root_dir)
+        if parts == (".",):
+            parts = ()
+        return parts
+
+    def _get_key_from_relative(self, path) -> Key:
+        parts = self.path.relparts(path, self.root_marker)
+        if parts and parts[0] == os.curdir:
+            parts = parts[1:]
+        return parts
+
+    def _from_key(self, parts: Key) -> str:
+        return self.repo.fs.path.join(self.repo.root_dir, *parts)
+
+    @property
+    def repo_url(self):
+        if self.repo is None:
+            return None
+        return self.repo.url
 
     @property
     def config(self):
-        raise NotImplementedError
-
-    def _get_key(self, path):
-        from dvc.fs.local import LocalFileSystem
-
-        from . import get_cloud_fs
-
-        cls, kwargs, fs_path = get_cloud_fs(None, url=path)
-
-        if cls != LocalFileSystem or os.path.isabs(path):
-            fs = cls(**kwargs)
-            return (cls.scheme, *fs.path.parts(fs_path))
-
-        fs_key = "repo"
-        key = path.split(self.sep)
-        if key == ["."] or key == [""]:
-            key = ()
-
-        return (fs_key, *key)
-
-    def _get_fs_path(self, path: "AnyPath", remote=None):
-        from dvc.config import NoRemoteError
-
-        info = self.info(path)
-        if info["type"] == "directory":
-            raise IsADirectoryError
-
-        value = info.get("md5")
-        if not value:
-            raise FileNotFoundError
-
-        cache_path = self.repo.odb.local.hash_to_path(value)
-
-        if self.repo.odb.local.fs.exists(cache_path):
-            return self.repo.odb.local.fs, cache_path
-
-        try:
-            remote_odb = self.repo.cloud.get_remote_odb(remote)
-        except NoRemoteError as exc:
-            raise FileNotFoundError from exc
-        remote_fs_path = remote_odb.hash_to_path(value)
-        return remote_odb.fs, remote_fs_path
-
-    def open(  # type: ignore
-        self, path: str, mode="r", encoding=None, **kwargs
-    ):  # pylint: disable=arguments-renamed, arguments-differ
-        fs, fspath = self._get_fs_path(path, **kwargs)
-        return fs.open(fspath, mode=mode, encoding=encoding)
-
-    def ls(self, path, detail=True, **kwargs):
-        info = self.info(path)
-        if info["type"] != "directory":
-            return [info] if detail else [path]
-
-        root_key = self._get_key(path)
-        try:
-            entries = [
-                self.sep.join((path, name)) if path else name
-                for name in self.repo.index.tree.ls(prefix=root_key)
-            ]
-        except KeyError as exc:
-            raise FileNotFoundError from exc
-
-        if not detail:
-            return entries
-
-        return [self.info(epath) for epath in entries]
-
-    def isdvc(self, path, recursive=False, strict=True):
-        try:
-            info = self.info(path)
-        except FileNotFoundError:
-            return False
-
-        recurse = recursive or not strict
-        return bool(info.get("outs") if recurse else info.get("isout"))
-
-    def info(self, path, **kwargs):
-        from dvc.data.meta import Meta
-
-        key = self._get_key(path)
-
-        try:
-            outs = list(self.repo.index.tree.iteritems(key))
-        except KeyError as exc:
-            raise FileNotFoundError from exc
-
-        ret = {
-            "type": "file",
-            "size": 0,
-            "isexec": False,
-            "isdvc": False,
-            "outs": outs,
-            "name": path,
+        return {
+            self.PARAM_REPO_URL: self.repo_url,
+            self.PARAM_REPO_ROOT: self.repo.root_dir,
+            self.PARAM_REV: getattr(self.repo.fs, "rev", None),
+            self.PARAM_CACHE_DIR: os.path.abspath(self.repo.odb.local.path),
+            self.PARAM_CACHE_TYPES: self.repo.odb.local.cache_types,
+            self.PARAM_SUBREPOS: self._traverse_subrepos,
         }
 
-        if len(outs) > 1 and outs[0][0] != key:
-            shortest = self.repo.index.tree.shortest_prefix(key)
-            if shortest:
-                assert shortest[1][1].isdir
-                if len(shortest[0]) <= len(key):
-                    ret["isdvc"] = True
+    @classmethod
+    def _repo_from_fs_config(
+        cls, **config
+    ) -> Tuple["Repo", Optional["RepoFactory"]]:
+        from dvc.external_repo import erepo_factory, external_repo
+        from dvc.repo import Repo
 
-            ret["type"] = "directory"
-            return ret
+        url = config.get(cls.PARAM_REPO_URL)
+        root = config.get(cls.PARAM_REPO_ROOT)
+        assert url or root
 
-        item_key, (meta, hash_info) = outs[0]
+        def _open(*args, **kwargs):
+            # NOTE: if original repo was an erepo (and has a URL),
+            # we cannot use Repo.open() since it will skip erepo
+            # cache/remote setup for local URLs
+            if url is None:
+                return Repo.open(*args, **kwargs)
+            return external_repo(*args, **kwargs)
 
-        meta = meta or Meta()
+        cache_dir = config.get(cls.PARAM_CACHE_DIR)
+        cache_config = (
+            {}
+            if not cache_dir
+            else {
+                "cache": {
+                    "dir": cache_dir,
+                    "type": config.get(cls.PARAM_CACHE_TYPES),
+                }
+            }
+        )
+        repo_kwargs: dict = {
+            "rev": config.get(cls.PARAM_REV),
+            "subrepos": config.get(cls.PARAM_SUBREPOS, False),
+            "uninitialized": True,
+        }
+        factory: Optional["RepoFactory"] = None
+        if url is None:
+            repo_kwargs["config"] = cache_config
+        else:
+            repo_kwargs["cache_dir"] = cache_dir
+            factory = erepo_factory(url, root, cache_config)
 
-        if key != item_key:
-            assert item_key[: len(key)] == key
-            ret["type"] = "directory"
-            return ret
+        with _open(
+            url if url else root,
+            **repo_kwargs,
+        ) as repo:
+            return repo, factory
 
-        ret["size"] = meta.size
-        ret["isexec"] = meta.isexec
-        ret[hash_info.name] = hash_info.value
-        ret["isdvc"] = True
-        ret["isout"] = True
-        ret["meta"] = meta
-        if hash_info and hash_info.isdir:
-            ret["type"] = "directory"
-        return ret
+    def _get_repo(self, key: Key) -> "Repo":
+        """Returns repo that the path falls in, using prefix.
 
-    def get_file(self, rpath, lpath, callback=DEFAULT_CALLBACK, **kwargs):
-        fs, path = self._get_fs_path(rpath)
-        fs.get_file(  # pylint: disable=protected-access
-            path, lpath, callback=callback, **kwargs
+        If the path is already tracked/collected, it just returns the repo.
+
+        Otherwise, it collects the repos that might be in the path's parents
+        and then returns the appropriate one.
+        """
+        repo = self._subrepos_trie.get(key)
+        if repo:
+            return repo
+
+        prefix_key, repo = self._subrepos_trie.longest_prefix(key)
+        prefix = self.repo.fs.path.join(
+            self.repo.root_dir,
+            *prefix_key,  # pylint: disable=not-an-iterable
         )
 
+        path = self._from_key(key)
+        parents = (parent for parent in self.repo.fs.path.parents(path))
+        dirs = [path] + list(takewhile(lambda p: p != prefix, parents))
+        dirs.reverse()
+        self._update(dirs, starting_repo=repo)
+        return self._subrepos_trie.get(key) or self.repo
+
+    @wrap_with(threading.Lock())
+    def _update(self, dirs, starting_repo):
+        """Checks for subrepo in directories and updates them."""
+        repo = starting_repo
+        for d in dirs:
+            key = self._get_key(d)
+            if self._is_dvc_repo(d):
+                repo = self.repo_factory(
+                    d,
+                    fs=self.repo.fs,
+                    scm=self.repo.scm,
+                    repo_factory=self.repo_factory,
+                )
+                self._datafss[key] = DataFileSystem(
+                    index=repo.index.data["repo"]
+                )
+            self._subrepos_trie[key] = repo
+
+    def _is_dvc_repo(self, dir_path):
+        """Check if the directory is a dvc repo."""
+        if not self._traverse_subrepos:
+            return False
+
+        from dvc.repo import Repo
+
+        repo_path = self.repo.fs.path.join(dir_path, Repo.DVC_DIR)
+        return self.repo.fs.isdir(repo_path)
+
+    def _get_fs_pair(
+        self, path
+    ) -> Tuple[
+        Optional["Repo"],
+        Optional[FileSystem],
+        Optional[str],
+        Optional[DataFileSystem],
+        Optional[str],
+    ]:
+        """
+        Returns a pair of fss based on repo the path falls in, using prefix.
+        """
+        key = self._get_key_from_relative(path)
+
+        fs_path = self._from_key(key)
+        repo = self._get_repo(key)
+        fs = repo.fs
+        if repo is self.repo:
+            dvc_parts = key
+            dvc_fs = self._datafss.get(())
+        else:
+            repo_parts = fs.path.relparts(repo.root_dir, self.repo.root_dir)
+            dvc_parts = key[len(repo_parts) :]
+            key = self._get_key(repo.root_dir)
+            dvc_fs = self._datafss.get(key)
+
+        if dvc_fs:
+            dvc_path = dvc_fs.path.join(*dvc_parts) if dvc_parts else ""
+        else:
+            dvc_path = None
+
+        return repo, fs, fs_path, dvc_fs, dvc_path
+
+    def open(
+        self, path, mode="r", encoding="utf-8", **kwargs
+    ):  # pylint: disable=arguments-renamed, arguments-differ
+        if "b" in mode:
+            encoding = None
+
+        _, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
+        try:
+            return fs.open(fs_path, mode=mode, encoding=encoding)
+        except FileNotFoundError:
+            if not dvc_fs:
+                raise
+
+        return dvc_fs.open(dvc_path, mode=mode, encoding=encoding, **kwargs)
+
+    def isdvc(self, path, **kwargs):
+        _, _, _, dvc_fs, dvc_path = self._get_fs_pair(path)
+        return dvc_fs is not None and dvc_fs.isdvc(dvc_path, **kwargs)
+
+    def ls(  # pylint: disable=arguments-differ
+        self, path, detail=True, dvc_only=False, **kwargs
+    ):
+        repo, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
+
+        dvcignore = repo.dvcignore
+        ignore_subrepos = kwargs.get("ignore_subrepos", True)
+
+        names = set()
+        if dvc_fs:
+            with suppress(FileNotFoundError):
+                for entry in dvc_fs.ls(dvc_path, detail=False):
+                    names.add(dvc_fs.path.name(entry))
+
+        if not dvc_only and fs:
+            try:
+                for entry in dvcignore.ls(
+                    fs, fs_path, detail=False, ignore_subrepos=ignore_subrepos
+                ):
+                    names.add(fs.path.name(entry))
+            except (FileNotFoundError, NotADirectoryError):
+                pass
+
+        dvcfiles = kwargs.get("dvcfiles", False)
+        if not dvcfiles:
+            names = (name for name in names if not _is_dvc_file(name))
+
+        infos = []
+        paths = []
+        for name in names:
+            entry_path = self.path.join(path, name)
+            try:
+                info = self.info(entry_path, ignore_subrepos=ignore_subrepos)
+            except FileNotFoundError:
+                continue
+            infos.append(info)
+            paths.append(entry_path)
+
+        if not detail:
+            return paths
+
+        return infos
+
+    def get_file(  # pylint: disable=arguments-differ
+        self, rpath, lpath, callback=DEFAULT_CALLBACK, **kwargs
+    ):
+        _, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(rpath)
+
+        if fs:
+            try:
+                fs.get_file(fs_path, lpath, callback=callback, **kwargs)
+                return
+            except FileNotFoundError:
+                if not dvc_fs:
+                    raise
+        dvc_fs.get_file(dvc_path, lpath, callback=callback, **kwargs)
+
+    def info(self, path, **kwargs):
+        repo, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
+
+        dvcignore = repo.dvcignore
+        ignore_subrepos = kwargs.get("ignore_subrepos", True)
+
+        dvc_info = None
+        if dvc_fs:
+            try:
+                dvc_info = dvc_fs.info(dvc_path)
+            except FileNotFoundError:
+                pass
+
+        fs_info = None
+        if fs:
+            try:
+                fs_info = fs.info(fs_path)
+                if dvcignore.is_ignored(
+                    fs, fs_path, ignore_subrepos=ignore_subrepos
+                ):
+                    fs_info = None
+            except (FileNotFoundError, NotADirectoryError):
+                if not dvc_info:
+                    raise
+
+        # NOTE: if some parent in fs_path turns out to be a file, it means
+        # that the whole repofs branch doesn't exist.
+        if fs and not fs_info and dvc_info:
+            for parent in fs.path.parents(fs_path):
+                try:
+                    if fs.info(parent)["type"] != "directory":
+                        dvc_info = None
+                        break
+                except FileNotFoundError:
+                    continue
+
+        if not dvc_info and not fs_info:
+            raise FileNotFoundError
+
+        info = _merge_info(repo, fs_info, dvc_info)
+        info["name"] = path
+        return info
+
     def checksum(self, path):
-        info = self.info(path)
-        md5 = info.get("md5")
-        if md5:
-            return md5
-        raise NotImplementedError
+        _, fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
+
+        try:
+            return fs.checksum(fs_path)
+        except FileNotFoundError:
+            return dvc_fs.checksum(dvc_path)
 
 
 class DvcFileSystem(FileSystem):
-    scheme = "local"
-
+    protocol = "local"
     PARAM_CHECKSUM = "md5"
 
     def _prepare_credentials(self, **config):
@@ -183,6 +455,24 @@ class DvcFileSystem(FileSystem):
     def isdvc(self, path, **kwargs):
         return self.fs.isdvc(path, **kwargs)
 
+    @property
+    def path(self):  # pylint: disable=invalid-overridden-method
+        return self.fs.path
+
     @property
     def repo(self):
         return self.fs.repo
+
+    @property
+    def repo_url(self):
+        return self.fs.repo_url
+
+    @property
+    def config(self):
+        return self.fs.config
+
+    def from_os_path(self, path):
+        if os.path.isabs(path):
+            path = os.path.relpath(path, self.repo.root_dir)
+
+        return as_posix(path)

dvc/fs/gdrive.py

@@ -1,259 +0,0 @@
-import logging
-import os
-import threading
-
-from funcy import cached_property, wrap_prop
-
-from dvc.exceptions import DvcException
-from dvc.scheme import Schemes
-from dvc.utils import format_link, tmp_fname
-
-from .base import FileSystem
-
-logger = logging.getLogger(__name__)
-FOLDER_MIME_TYPE = "application/vnd.google-apps.folder"
-
-
-class GDriveAuthError(DvcException):
-    def __init__(self, cred_location):
-
-        if cred_location:
-            message = (
-                "GDrive remote auth failed with credentials in '{}'.\n"
-                "Backup first, remove or fix them, and run DVC again.\n"
-                "It should do auth again and refresh the credentials.\n\n"
-                "Details:".format(cred_location)
-            )
-        else:
-            message = "Failed to authenticate GDrive remote"
-
-        super().__init__(message)
-
-
-class GDriveFileSystem(FileSystem):  # pylint:disable=abstract-method
-    scheme = Schemes.GDRIVE
-    PARAM_CHECKSUM = "checksum"
-    REQUIRES = {"pydrive2": "pydrive2"}
-    # Always prefer traverse for GDrive since API usage quotas are a concern.
-    TRAVERSE_WEIGHT_MULTIPLIER = 1
-
-    GDRIVE_CREDENTIALS_DATA = "GDRIVE_CREDENTIALS_DATA"
-    DEFAULT_USER_CREDENTIALS_FILE = "gdrive-user-credentials.json"
-
-    DEFAULT_GDRIVE_CLIENT_ID = "710796635688-iivsgbgsb6uv1fap6635dhvuei09o66c.apps.googleusercontent.com"  # noqa: E501
-    DEFAULT_GDRIVE_CLIENT_SECRET = "a1Fz59uTpVNeG_VGuSKDLJXv"
-
-    def __init__(self, **config):
-        from fsspec.utils import infer_storage_options
-
-        super().__init__(**config)
-
-        self.url = config["url"]
-        opts = infer_storage_options(self.url)
-
-        if not opts["host"]:
-            raise DvcException(
-                "Empty GDrive URL '{}'. Learn more at {}".format(
-                    config["url"],
-                    format_link("https://man.dvc.org/remote/add"),
-                )
-            )
-
-        self._bucket = opts["host"]
-        self._path = opts["path"].lstrip("/")
-        self._trash_only = config.get("gdrive_trash_only")
-        self._use_service_account = config.get("gdrive_use_service_account")
-        self._service_account_user_email = config.get(
-            "gdrive_service_account_user_email"
-        )
-        self._service_account_json_file_path = config.get(
-            "gdrive_service_account_json_file_path"
-        )
-        self._client_id = config.get("gdrive_client_id")
-        self._client_secret = config.get("gdrive_client_secret")
-        self._validate_config()
-
-        tmp_dir = config["gdrive_credentials_tmp_dir"]
-        assert tmp_dir
-
-        self._gdrive_service_credentials_path = tmp_fname(
-            os.path.join(tmp_dir, "")
-        )
-        self._gdrive_user_credentials_path = (
-            tmp_fname(os.path.join(tmp_dir, ""))
-            if os.getenv(GDriveFileSystem.GDRIVE_CREDENTIALS_DATA)
-            else config.get(
-                "gdrive_user_credentials_file",
-                os.path.join(tmp_dir, self.DEFAULT_USER_CREDENTIALS_FILE),
-            )
-        )
-
-    @classmethod
-    def _strip_protocol(cls, path):
-        from fsspec.utils import infer_storage_options
-
-        opts = infer_storage_options(path)
-
-        return "{host}{path}".format(**opts)
-
-    def unstrip_protocol(self, path):
-        return f"gdrive://{path}"
-
-    @staticmethod
-    def _get_kwargs_from_urls(urlpath):
-        return {"url": urlpath}
-
-    def _validate_config(self):
-        # Validate Service Account configuration
-        if (
-            self._use_service_account
-            and not self._service_account_json_file_path
-            and GDriveFileSystem.GDRIVE_CREDENTIALS_DATA not in os.environ
-        ):
-            raise DvcException(
-                "To use service account, set "
-                "`gdrive_service_account_json_file_path`, and optionally"
-                "`gdrive_service_account_user_email` in DVC config\n"
-                "Learn more at {}".format(
-                    format_link("https://man.dvc.org/remote/modify")
-                )
-            )
-
-        # Validate OAuth 2.0 Client ID configuration
-        if not self._use_service_account:
-            if bool(self._client_id) != bool(self._client_secret):
-                raise DvcException(
-                    "Please specify GDrive's client ID and secret in "
-                    "DVC config or omit both to use the defaults.\n"
-                    "Learn more at {}".format(
-                        format_link("https://man.dvc.org/remote/modify")
-                    )
-                )
-
-    @cached_property
-    def credentials_location(self):
-        """
-        Helper to determine where will GDrive remote read credentials from.
-        Useful for tests, exception messages, etc. Returns either env variable
-        name if it's set or actual path to the credentials file.
-        """
-        if os.getenv(GDriveFileSystem.GDRIVE_CREDENTIALS_DATA):
-            return GDriveFileSystem.GDRIVE_CREDENTIALS_DATA
-        if os.path.exists(self._gdrive_user_credentials_path):
-            return self._gdrive_user_credentials_path
-        return None
-
-    @staticmethod
-    def _validate_credentials(auth, settings):
-        """
-        Detects discrepancy in DVC config and cached credentials file.
-        Usually happens when a second remote is added and it is using
-        the same credentials default file. Or when someones decides to change
-        DVC config client id or secret but forgets to remove the cached
-        credentials file.
-        """
-        if not os.getenv(GDriveFileSystem.GDRIVE_CREDENTIALS_DATA):
-            if (
-                settings["client_config"]["client_id"]
-                != auth.credentials.client_id
-                or settings["client_config"]["client_secret"]
-                != auth.credentials.client_secret
-            ):
-                logger.warning(
-                    "Client ID and secret configured do not match the "
-                    "actual ones used\nto access the remote. Do you "
-                    "use multiple GDrive remotes and forgot to\nset "
-                    "`gdrive_user_credentials_file` for one or more of them? "
-                    "Learn more at\n{}.\n".format(
-                        format_link("https://man.dvc.org/remote/modify")
-                    )
-                )
-
-    @wrap_prop(threading.RLock())
-    @cached_property
-    def fs(self):
-        from pydrive2.auth import GoogleAuth
-        from pydrive2.fs import GDriveFileSystem as _GDriveFileSystem
-
-        temporary_save_path = self._gdrive_user_credentials_path
-        is_credentials_temp = os.getenv(
-            GDriveFileSystem.GDRIVE_CREDENTIALS_DATA
-        )
-        if self._use_service_account:
-            temporary_save_path = self._gdrive_service_credentials_path
-
-        if is_credentials_temp:
-            with open(temporary_save_path, "w", encoding="utf-8") as cred_file:
-                cred_file.write(
-                    os.getenv(GDriveFileSystem.GDRIVE_CREDENTIALS_DATA)
-                )
-
-        auth_settings = {
-            "client_config_backend": "settings",
-            "save_credentials": True,
-            "save_credentials_backend": "file",
-            "save_credentials_file": self._gdrive_user_credentials_path,
-            "get_refresh_token": True,
-            "oauth_scope": [
-                "https://www.googleapis.com/auth/drive",
-                "https://www.googleapis.com/auth/drive.appdata",
-            ],
-        }
-
-        if self._use_service_account:
-            auth_settings["service_config"] = {
-                "client_user_email": self._service_account_user_email,
-                "client_json_file_path": self._service_account_json_file_path,
-            }
-            if is_credentials_temp:
-                auth_settings["service_config"][
-                    "client_json_file_path"
-                ] = temporary_save_path
-
-        else:
-            auth_settings["client_config"] = {
-                "client_id": self._client_id or self.DEFAULT_GDRIVE_CLIENT_ID,
-                "client_secret": self._client_secret
-                or self.DEFAULT_GDRIVE_CLIENT_SECRET,
-                "auth_uri": "https://accounts.google.com/o/oauth2/auth",
-                "token_uri": "https://oauth2.googleapis.com/token",
-                "revoke_uri": "https://oauth2.googleapis.com/revoke",
-                "redirect_uri": "",
-            }
-
-        GoogleAuth.DEFAULT_SETTINGS.update(auth_settings)
-
-        # Pass non existent settings path to force DEFAULT_SETTINGS loadings
-        gauth = GoogleAuth(settings_file="")
-
-        try:
-            logger.debug(
-                "GDrive remote auth with config '{}'.".format(
-                    GoogleAuth.DEFAULT_SETTINGS
-                )
-            )
-            if self._use_service_account:
-                gauth.ServiceAuth()
-            else:
-                gauth.CommandLineAuth()
-                GDriveFileSystem._validate_credentials(gauth, auth_settings)
-
-        # Handle AuthenticationError, RefreshError and other auth failures
-        # It's hard to come up with a narrow exception, since PyDrive throws
-        # a lot of different errors - broken credentials file, refresh token
-        # expired, flow failed, etc.
-        except Exception as exc:
-            raise GDriveAuthError(self.credentials_location) from exc
-        finally:
-            if is_credentials_temp:
-                os.remove(temporary_save_path)
-
-        return _GDriveFileSystem(
-            f"{self._bucket}/{self._path}",
-            gauth,
-            trash_only=self._trash_only,
-        )
-
-    def upload_fobj(self, fobj, to_info, **kwargs):
-        self.makedirs(os.path.dirname(to_info))
-        return self.fs.upload_fobj(fobj, to_info, **kwargs)

dvc/fs/git.py

@@ -1,10 +1,9 @@
-import os
 import threading
 from typing import TYPE_CHECKING, Any
 
 from funcy import cached_property, wrap_prop
 
-from .base import FileSystem
+from . import FileSystem
 
 if TYPE_CHECKING:
     from scmrepo.fs import GitFileSystem as FsspecGitFileSystem
@@ -15,8 +14,7 @@ if TYPE_CHECKING:
 class GitFileSystem(FileSystem):  # pylint:disable=abstract-method
     """Proxies the repo file access methods to Git objects"""
 
-    sep = os.sep
-    scheme = "local"
+    protocol = "local"
 
     def __init__(
         self,
@@ -47,6 +45,10 @@ class GitFileSystem(FileSystem):  # pylint:disable=abstract-method
 
         return FsspecGitFileSystem(**self.fs_args)
 
+    @cached_property
+    def path(self):
+        return self.fs.path
+
     @property
     def rev(self) -> str:
         return self.fs.rev

dvc/fs/gs.py

@@ -1,27 +0,0 @@
-import threading
-
-from funcy import cached_property, wrap_prop
-
-from dvc.scheme import Schemes
-
-# pylint:disable=abstract-method
-from .base import ObjectFileSystem
-
-
-class GSFileSystem(ObjectFileSystem):
-    scheme = Schemes.GS
-    REQUIRES = {"gcsfs": "gcsfs"}
-    PARAM_CHECKSUM = "etag"
-
-    def _prepare_credentials(self, **config):
-        login_info = {"consistency": None}
-        login_info["project"] = config.get("projectname")
-        login_info["token"] = config.get("credentialpath")
-        return login_info
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        from gcsfs import GCSFileSystem
-
-        return GCSFileSystem(**self.fs_args)

dvc/fs/hdfs.py

@@ -1,99 +0,0 @@
-import os
-import re
-import subprocess
-import threading
-
-from funcy import cached_property, wrap_prop
-
-from dvc.scheme import Schemes
-from dvc.utils import fix_env
-
-from .base import FileSystem
-
-CHECKSUM_REGEX = re.compile(r".*\t.*\t(?P<checksum>.*)")
-
-
-# pylint: disable=abstract-method
-class HDFSFileSystem(FileSystem):
-    scheme = Schemes.HDFS
-    REQUIRES = {"fsspec": "fsspec", "pyarrow": "pyarrow"}
-    PARAM_CHECKSUM = "checksum"
-
-    @classmethod
-    def _strip_protocol(cls, path: str) -> str:
-        from fsspec.utils import infer_storage_options
-
-        return infer_storage_options(path)["path"]
-
-    def unstrip_protocol(self, path: str) -> str:
-        host = self.fs_args["host"]
-        port = self.fs_args.get("port")
-        netloc = host + (f":{port}" if port else "")
-        return "hdfs://" + netloc + "/" + path.lstrip("/")
-
-    @staticmethod
-    def _get_kwargs_from_urls(urlpath):
-        from fsspec.implementations.arrow import HadoopFileSystem
-
-        # pylint:disable=protected-access
-        return HadoopFileSystem._get_kwargs_from_urls(urlpath)
-
-    def _prepare_credentials(self, **config):
-        return config
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        from fsspec.implementations.arrow import HadoopFileSystem
-
-        return HadoopFileSystem(**self.fs_args)
-
-    def checksum(self, path):
-        return self._checksum(path)
-
-    def _checksum(self, path, **kwargs):
-        # PyArrow doesn't natively support retrieving the
-        # checksum, so we have to use hadoop fs
-
-        url = self.unstrip_protocol(path)
-
-        result = self._run_command(
-            f"checksum {url}",
-            env=fix_env(os.environ),
-            user=self.fs_args.get("user"),
-        )
-        if result is None:
-            return None
-
-        match = CHECKSUM_REGEX.match(result)
-        if match is None:
-            return None
-
-        return match.group("checksum")
-
-    def _run_command(self, cmd, env=None, user=None):
-        cmd = "hadoop fs -" + cmd
-        if user:
-            cmd = f"HADOOP_USER_NAME={user} " + cmd
-
-        # NOTE: close_fds doesn't work with redirected stdin/stdout/stderr.
-        # See https://github.com/iterative/dvc/issues/1197.
-        close_fds = os.name != "nt"
-
-        executable = os.getenv("SHELL") if os.name != "nt" else None
-        p = subprocess.Popen(
-            cmd,
-            shell=True,
-            close_fds=close_fds,
-            executable=executable,
-            env=env or os.environ,
-            stdin=subprocess.PIPE,
-            stdout=subprocess.PIPE,
-            stderr=subprocess.PIPE,
-        )
-        out, err = p.communicate()
-
-        if p.returncode != 0:
-            raise subprocess.CalledProcessError(p.returncode, cmd, out, err)
-        else:
-            return out.decode("utf-8")

dvc/fs/http.py

@@ -1,172 +0,0 @@
-import threading
-from typing import IO, Union
-
-from funcy import cached_property, memoize, wrap_with
-
-from dvc import prompt
-from dvc.scheme import Schemes
-
-from ._callback import DEFAULT_CALLBACK, FsspecCallback
-from .base import AnyFSPath, FileSystem
-
-
-@wrap_with(threading.Lock())
-@memoize
-def ask_password(host, user):
-    return prompt.password(
-        "Enter a password for "
-        "host '{host}' user '{user}'".format(host=host, user=user)
-    )
-
-
-def make_context(ssl_verify):
-    if isinstance(ssl_verify, bool) or ssl_verify is None:
-        return ssl_verify
-
-    # If this is a path, then we will create an
-    # SSL context for it, and load the given certificate.
-    import ssl
-
-    context = ssl.create_default_context()
-    context.load_verify_locations(ssl_verify)
-    return context
-
-
-# pylint: disable=abstract-method
-class HTTPFileSystem(FileSystem):
-    scheme = Schemes.HTTP
-    PARAM_CHECKSUM = "checksum"
-    REQUIRES = {"aiohttp": "aiohttp", "aiohttp-retry": "aiohttp_retry"}
-    CAN_TRAVERSE = False
-
-    SESSION_RETRIES = 5
-    SESSION_BACKOFF_FACTOR = 0.1
-    REQUEST_TIMEOUT = 60
-
-    def _prepare_credentials(self, **config):
-        import aiohttp
-        from fsspec.asyn import fsspec_loop
-
-        from dvc.config import ConfigError
-
-        credentials = {}
-        client_kwargs = credentials.setdefault("client_kwargs", {})
-
-        if config.get("auth"):
-            user = config.get("user")
-            password = config.get("password")
-            custom_auth_header = config.get("custom_auth_header")
-
-            if password is None and config.get("ask_password"):
-                password = ask_password(config.get("url"), user or "custom")
-
-            auth_method = config["auth"]
-            if auth_method == "basic":
-                if user is None or password is None:
-                    raise ConfigError(
-                        "HTTP 'basic' authentication require both "
-                        "'user' and 'password'"
-                    )
-
-                client_kwargs["auth"] = aiohttp.BasicAuth(user, password)
-            elif auth_method == "custom":
-                if custom_auth_header is None or password is None:
-                    raise ConfigError(
-                        "HTTP 'custom' authentication require both "
-                        "'custom_auth_header' and 'password'"
-                    )
-                credentials["headers"] = {custom_auth_header: password}
-            else:
-                raise NotImplementedError(
-                    f"Auth method {auth_method!r} is not supported."
-                )
-
-        # Force cleanup of closed SSL transports.
-        # https://github.com/iterative/dvc/issues/7414
-        connector_kwargs = {"enable_cleanup_closed": True}
-
-        if "ssl_verify" in config:
-            connector_kwargs.update(ssl=make_context(config["ssl_verify"]))
-
-        with fsspec_loop():
-            client_kwargs["connector"] = aiohttp.TCPConnector(
-                **connector_kwargs
-            )
-        # The connector should not be owned by aiohttp.ClientSession since
-        # it is closed by fsspec (HTTPFileSystem.close_session)
-        client_kwargs["connector_owner"] = False
-
-        # Allow reading proxy configurations from the environment.
-        client_kwargs["trust_env"] = True
-
-        credentials["get_client"] = self.get_client
-        self.upload_method = config.get("method", "POST")
-        return credentials
-
-    async def get_client(self, **kwargs):
-        import aiohttp
-        from aiohttp_retry import ExponentialRetry, RetryClient
-
-        kwargs["retry_options"] = ExponentialRetry(
-            attempts=self.SESSION_RETRIES,
-            factor=self.SESSION_BACKOFF_FACTOR,
-            max_timeout=self.REQUEST_TIMEOUT,
-            exceptions=[aiohttp.ClientError],
-        )
-
-        # The default timeout for the aiohttp is 300 seconds
-        # which is too low for DVC's interactions (especially
-        # on the read) when dealing with large data blobs. We
-        # unlimit the total time to read, and only limit the
-        # time that is spent when connecting to the remote server.
-        kwargs["timeout"] = aiohttp.ClientTimeout(
-            total=None,
-            connect=self.REQUEST_TIMEOUT,
-            sock_connect=self.REQUEST_TIMEOUT,
-            sock_read=self.REQUEST_TIMEOUT,
-        )
-
-        return RetryClient(**kwargs)
-
-    @cached_property
-    def fs(self):
-        from fsspec.implementations.http import (
-            HTTPFileSystem as _HTTPFileSystem,
-        )
-
-        return _HTTPFileSystem(**self.fs_args)
-
-    def unstrip_protocol(self, path: str) -> str:
-        return path
-
-    def put_file(
-        self,
-        from_file: Union[AnyFSPath, IO],
-        to_info: AnyFSPath,
-        callback: FsspecCallback = DEFAULT_CALLBACK,
-        size: int = None,
-        **kwargs,
-    ) -> None:
-        kwargs.setdefault("method", self.upload_method)
-        super().put_file(
-            from_file, to_info, callback=callback, size=size, **kwargs
-        )
-
-    # pylint: disable=arguments-differ
-
-    def find(self, *args, **kwargs):
-        raise NotImplementedError
-
-    def isdir(self, *args, **kwargs):
-        return False
-
-    def isfile(self, *args, **kwargs):
-        return True
-
-    def ls(self, *args, **kwargs):
-        raise NotImplementedError
-
-    def walk(self, *args, **kwargs):
-        raise NotImplementedError
-
-    # pylint: enable=arguments-differ

dvc/fs/https.py

@@ -1,7 +0,0 @@
-from dvc.scheme import Schemes
-
-from .http import HTTPFileSystem
-
-
-class HTTPSFileSystem(HTTPFileSystem):  # pylint:disable=abstract-method
-    scheme = Schemes.HTTPS

dvc/fs/local.py

@@ -1,189 +0,0 @@
-import logging
-import os
-import threading
-
-from fsspec import AbstractFileSystem
-from funcy import cached_property, wrap_prop
-
-from dvc.scheme import Schemes
-from dvc.system import System
-from dvc.utils import tmp_fname
-from dvc.utils.fs import copy_fobj_to_file, copyfile, makedirs, move, remove
-
-from .base import FileSystem
-
-logger = logging.getLogger(__name__)
-
-
-# pylint:disable=abstract-method, arguments-differ
-class FsspecLocalFileSystem(AbstractFileSystem):
-    sep = os.sep
-
-    def __init__(self, *args, **kwargs):
-        from fsspec.implementations.local import LocalFileSystem as LocalFS
-
-        super().__init__(*args, **kwargs)
-        self.fs = LocalFS()
-
-    def makedirs(self, path, exist_ok=False):
-        makedirs(path, exist_ok=exist_ok)
-
-    def mkdir(self, path, create_parents=True, **kwargs):
-        if self.exists(path):
-            raise FileExistsError(path)
-        if create_parents:
-            self.makedirs(path, exist_ok=True)
-        else:
-            os.mkdir(path, **kwargs)
-
-    def lexists(self, path, **kwargs):
-        return os.path.lexists(path)
-
-    def exists(self, path, **kwargs):
-        # TODO: replace this with os.path.exists once the problem is fixed on
-        # the fsspec https://github.com/intake/filesystem_spec/issues/742
-        return os.path.lexists(path)
-
-    def checksum(self, path) -> str:
-        from fsspec.utils import tokenize
-
-        st = os.stat(path)
-        return str(int(tokenize([st.st_ino, st.st_mtime, st.st_size]), 16))
-
-    def info(self, path, **kwargs):
-        return self.fs.info(path)
-
-    def ls(self, path, **kwargs):
-        return self.fs.ls(path, **kwargs)
-
-    def isfile(self, path) -> bool:
-        return os.path.isfile(path)
-
-    def isdir(self, path) -> bool:
-        return os.path.isdir(path)
-
-    def walk(self, path, maxdepth=None, topdown=True, **kwargs):
-        """Directory fs generator.
-
-        See `os.walk` for the docs. Differences:
-        - no support for symlinks
-        """
-        for root, dirs, files in os.walk(
-            path,
-            topdown=topdown,
-        ):
-            yield os.path.normpath(root), dirs, files
-
-    def find(self, path, **kwargs):
-        for root, _, files in self.walk(path, **kwargs):
-            for file in files:
-                # NOTE: os.path.join is ~5.5 times slower
-                yield f"{root}{os.sep}{file}"
-
-    @classmethod
-    def _parent(cls, path):
-        return os.path.dirname(path)
-
-    def put_file(self, lpath, rpath, callback=None, **kwargs):
-        parent = self._parent(rpath)
-        makedirs(parent, exist_ok=True)
-        tmp_file = os.path.join(parent, tmp_fname())
-        copyfile(lpath, tmp_file, callback=callback)
-        os.replace(tmp_file, rpath)
-
-    def get_file(self, rpath, lpath, callback=None, **kwargs):
-        copyfile(rpath, lpath, callback=callback)
-
-    def mv(self, path1, path2, **kwargs):
-        self.makedirs(self._parent(path2), exist_ok=True)
-        move(path1, path2)
-
-    def rmdir(self, path):
-        os.rmdir(path)
-
-    def rm_file(self, path):
-        remove(path)
-
-    def rm(self, path, recursive=False, maxdepth=None):
-        remove(path)
-
-    def copy(self, path1, path2, recursive=False, on_error=None, **kwargs):
-        tmp_info = os.path.join(self._parent(path2), tmp_fname(""))
-        try:
-            copyfile(path1, tmp_info)
-            os.rename(tmp_info, path2)
-        except Exception:
-            self.rm_file(tmp_info)
-            raise
-
-    def open(self, path, mode="r", encoding=None, **kwargs):
-        return open(path, mode=mode, encoding=encoding)
-
-    def symlink(self, path1, path2):
-        return System.symlink(path1, path2)
-
-    @staticmethod
-    def is_symlink(path):
-        return System.is_symlink(path)
-
-    @staticmethod
-    def is_hardlink(path):
-        return System.is_hardlink(path)
-
-    def hardlink(self, path1, path2):
-        # If there are a lot of empty files (which happens a lot in datasets),
-        # and the cache type is `hardlink`, we might reach link limits and
-        # will get something like: `too many links error`
-        #
-        # This is because all those empty files will have the same hash
-        # (i.e. 68b329da9893e34099c7d8ad5cb9c940), therefore, they will be
-        # linked to the same file in the cache.
-        #
-        # From https://en.wikipedia.org/wiki/Hard_link
-        #   * ext4 limits the number of hard links on a file to 65,000
-        #   * Windows with NTFS has a limit of 1024 hard links on a file
-        #
-        # That's why we simply create an empty file rather than a link.
-        if self.size(path1) == 0:
-            self.open(path2, "w").close()
-
-            logger.debug("Created empty file: %s -> %s", path1, path2)
-            return
-
-        return System.hardlink(path1, path2)
-
-    def reflink(self, path1, path2):
-        return System.reflink(path1, path2)
-
-
-class LocalFileSystem(FileSystem):
-    sep = os.sep
-
-    scheme = Schemes.LOCAL
-    PARAM_CHECKSUM = "md5"
-    PARAM_PATH = "path"
-    TRAVERSE_PREFIX_LEN = 2
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        return FsspecLocalFileSystem(**self.config)
-
-    @cached_property
-    def path(self):
-        from .path import Path
-
-        return Path(self.sep, os.getcwd)
-
-    def upload_fobj(self, fobj, to_info, **kwargs):
-        self.makedirs(self.path.parent(to_info))
-        tmp_info = self.path.join(self.path.parent(to_info), tmp_fname(""))
-        try:
-            copy_fobj_to_file(fobj, tmp_info)
-            os.rename(tmp_info, to_info)
-        except Exception:
-            self.remove(tmp_info)
-            raise
-
-
-localfs = LocalFileSystem()

dvc/fs/memory.py

@@ -1,26 +0,0 @@
-import threading
-
-from funcy import cached_property, wrap_prop
-
-from dvc.scheme import Schemes
-
-from .base import FileSystem
-
-
-class MemoryFileSystem(FileSystem):  # pylint:disable=abstract-method
-    scheme = Schemes.MEMORY
-    PARAM_CHECKSUM = "md5"
-
-    def __eq__(self, other):
-        # NOTE: all fsspec MemoryFileSystem instances are equivalent and use a
-        # single global store
-        return isinstance(other, type(self))
-
-    __hash__ = FileSystem.__hash__
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        from fsspec.implementations.memory import MemoryFileSystem as MemFS
-
-        return MemFS(**self.fs_args)

dvc/fs/oss.py

@@ -1,37 +0,0 @@
-import logging
-import os
-import threading
-
-from funcy import cached_property, wrap_prop
-
-from dvc.scheme import Schemes
-
-from .base import ObjectFileSystem
-
-logger = logging.getLogger(__name__)
-
-
-# pylint:disable=abstract-method
-class OSSFileSystem(ObjectFileSystem):
-    scheme = Schemes.OSS
-    REQUIRES = {"ossfs": "ossfs"}
-    PARAM_CHECKSUM = "etag"
-    LIST_OBJECT_PAGE_SIZE = 100
-
-    def _prepare_credentials(self, **config):
-        login_info = {}
-        login_info["key"] = config.get("oss_key_id") or os.getenv(
-            "OSS_ACCESS_KEY_ID"
-        )
-        login_info["secret"] = config.get("oss_key_secret") or os.getenv(
-            "OSS_ACCESS_KEY_SECRET"
-        )
-        login_info["endpoint"] = config.get("oss_endpoint")
-        return login_info
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        from ossfs import OSSFileSystem as _OSSFileSystem
-
-        return _OSSFileSystem(**self.fs_args)

dvc/fs/path.py

@@ -1,114 +0,0 @@
-import ntpath
-import posixpath
-
-
-class Path:
-    def __init__(self, sep, getcwd=None):
-        def _getcwd():
-            return ""
-
-        self.getcwd = getcwd or _getcwd
-        if sep == posixpath.sep:
-            self.flavour = posixpath
-        elif sep == ntpath.sep:
-            self.flavour = ntpath
-        else:
-            raise ValueError(f"unsupported separator '{sep}'")
-
-    def join(self, *parts):
-        return self.flavour.join(*parts)
-
-    def split(self, path):
-        return self.flavour.split(path)
-
-    def normpath(self, path):
-        return self.flavour.normpath(path)
-
-    def isabs(self, path):
-        return self.flavour.isabs(path)
-
-    def abspath(self, path):
-        if not self.isabs(path):
-            path = self.join(self.getcwd(), path)
-        return self.normpath(path)
-
-    def commonprefix(self, path):
-        return self.flavour.commonprefix(path)
-
-    def parts(self, path):
-        drive, path = self.flavour.splitdrive(path.rstrip(self.flavour.sep))
-
-        ret = []
-        while True:
-            path, part = self.flavour.split(path)
-
-            if part:
-                ret.append(part)
-                continue
-
-            if path:
-                ret.append(path)
-
-            break
-
-        ret.reverse()
-
-        if drive:
-            ret = [drive] + ret
-
-        return tuple(ret)
-
-    def parent(self, path):
-        return self.flavour.dirname(path)
-
-    def dirname(self, path):
-        return self.parent(path)
-
-    def parents(self, path):
-        parts = self.parts(path)
-        return tuple(
-            self.join(*parts[:length])
-            for length in range(len(parts) - 1, 0, -1)
-        )
-
-    def name(self, path):
-        return self.parts(path)[-1]
-
-    def suffix(self, path):
-        name = self.name(path)
-        _, dot, suffix = name.partition(".")
-        return dot + suffix
-
-    def with_name(self, path, name):
-        parts = list(self.parts(path))
-        parts[-1] = name
-        return self.join(*parts)
-
-    def with_suffix(self, path, suffix):
-        parts = list(self.parts(path))
-        real_path, _, _ = parts[-1].partition(".")
-        parts[-1] = real_path + suffix
-        return self.join(*parts)
-
-    def isin(self, left, right):
-        left_parts = self.parts(left)
-        right_parts = self.parts(right)
-        left_len = len(left_parts)
-        right_len = len(right_parts)
-        return left_len > right_len and left_parts[:right_len] == right_parts
-
-    def isin_or_eq(self, left, right):
-        return left == right or self.isin(left, right)
-
-    def overlaps(self, left, right):
-        # pylint: disable=arguments-out-of-order
-        return self.isin_or_eq(left, right) or self.isin(right, left)
-
-    def relpath(self, path, start):
-        return self.flavour.relpath(path, start=start)
-
-    def relparts(self, path, base):
-        return self.parts(self.relpath(path, base))
-
-    def as_posix(self, path):
-        return path.replace(self.flavour.sep, posixpath.sep)

dvc/fs/repo.py

@@ -1,463 +0,0 @@
-import logging
-import os
-import threading
-from contextlib import suppress
-from itertools import takewhile
-from typing import TYPE_CHECKING, Callable, Optional, Tuple, Type, Union
-
-from fsspec.spec import AbstractFileSystem
-from funcy import cached_property, wrap_prop, wrap_with
-
-from ._callback import DEFAULT_CALLBACK
-from .base import FileSystem
-from .dvc import DvcFileSystem
-from .path import Path
-
-if TYPE_CHECKING:
-    from dvc.repo import Repo
-
-logger = logging.getLogger(__name__)
-
-RepoFactory = Union[Callable[[str], "Repo"], Type["Repo"]]
-
-
-def _wrap_walk(dvc_fs, *args, **kwargs):
-    for root, dnames, fnames in dvc_fs.walk(*args, **kwargs):
-        yield dvc_fs.path.join(dvc_fs.repo.root_dir, root), dnames, fnames
-
-
-def _ls(fs, path):
-    dnames = []
-    fnames = []
-
-    for entry in fs.ls(path, detail=True):
-        name = fs.path.name(entry["name"])
-        if entry["type"] == "directory":
-            dnames.append(name)
-        else:
-            fnames.append(name)
-
-    return dnames, fnames
-
-
-def _merge_info(repo, fs_info, dvc_info):
-    from dvc.utils import is_exec
-
-    ret = {"repo": repo}
-
-    if dvc_info:
-        ret["dvc_info"] = dvc_info
-        ret["type"] = dvc_info["type"]
-        ret["size"] = dvc_info["size"]
-        if not fs_info and "md5" in dvc_info:
-            ret["md5"] = dvc_info["md5"]
-
-    if fs_info:
-        ret["type"] = fs_info["type"]
-        ret["size"] = fs_info["size"]
-        isexec = False
-        if fs_info["type"] == "file":
-            isexec = is_exec(fs_info["mode"])
-        ret["isexec"] = isexec
-
-    return ret
-
-
-class _RepoFileSystem(AbstractFileSystem):  # pylint:disable=abstract-method
-    """DVC + git-tracked files fs.
-
-    Args:
-        repo: DVC or git repo.
-        subrepos: traverse to subrepos (by default, it ignores subrepos)
-        repo_factory: A function to initialize subrepo with, default is Repo.
-        kwargs: Additional keyword arguments passed to the `DvcFileSystem()`.
-    """
-
-    PARAM_REPO_URL = "repo_url"
-    PARAM_REPO_ROOT = "repo_root"
-    PARAM_REV = "rev"
-    PARAM_CACHE_DIR = "cache_dir"
-    PARAM_CACHE_TYPES = "cache_types"
-    PARAM_SUBREPOS = "subrepos"
-
-    def __init__(
-        self,
-        repo: Optional["Repo"] = None,
-        subrepos=False,
-        repo_factory: RepoFactory = None,
-        **kwargs,
-    ):
-        super().__init__()
-
-        from pygtrie import Trie
-
-        if repo is None:
-            repo, repo_factory = self._repo_from_fs_config(
-                subrepos=subrepos, **kwargs
-            )
-
-        if not repo_factory:
-            from dvc.repo import Repo
-
-            self.repo_factory: RepoFactory = Repo
-        else:
-            self.repo_factory = repo_factory
-
-        self.path = Path(self.sep)
-        self.repo = repo
-        self.hash_jobs = repo.fs.hash_jobs
-        self._traverse_subrepos = subrepos
-
-        self._subrepos_trie = Trie()
-        """Keeps track of each and every path with the corresponding repo."""
-
-        key = self._get_key(self.repo.root_dir)
-        self._subrepos_trie[key] = repo
-
-        self._dvcfss = {}
-        """Keep a dvcfs instance of each repo."""
-
-        if hasattr(repo, "dvc_dir"):
-            self._dvcfss[key] = DvcFileSystem(repo=repo)
-
-    def _get_key(self, path):
-        parts = self.repo.fs.path.relparts(path, self.repo.root_dir)
-        if parts == (".",):
-            parts = ()
-        return parts
-
-    @property
-    def repo_url(self):
-        if self.repo is None:
-            return None
-        return self.repo.url
-
-    @property
-    def config(self):
-        return {
-            self.PARAM_REPO_URL: self.repo_url,
-            self.PARAM_REPO_ROOT: self.repo.root_dir,
-            self.PARAM_REV: getattr(self.repo.fs, "rev", None),
-            self.PARAM_CACHE_DIR: os.path.abspath(
-                self.repo.odb.local.cache_dir
-            ),
-            self.PARAM_CACHE_TYPES: self.repo.odb.local.cache_types,
-            self.PARAM_SUBREPOS: self._traverse_subrepos,
-        }
-
-    @classmethod
-    def _repo_from_fs_config(
-        cls, **config
-    ) -> Tuple["Repo", Optional["RepoFactory"]]:
-        from dvc.external_repo import erepo_factory, external_repo
-        from dvc.repo import Repo
-
-        url = config.get(cls.PARAM_REPO_URL)
-        root = config.get(cls.PARAM_REPO_ROOT)
-        assert url or root
-
-        def _open(*args, **kwargs):
-            # NOTE: if original repo was an erepo (and has a URL),
-            # we cannot use Repo.open() since it will skip erepo
-            # cache/remote setup for local URLs
-            if url is None:
-                return Repo.open(*args, **kwargs)
-            return external_repo(*args, **kwargs)
-
-        cache_dir = config.get(cls.PARAM_CACHE_DIR)
-        cache_config = (
-            {}
-            if not cache_dir
-            else {
-                "cache": {
-                    "dir": cache_dir,
-                    "type": config.get(cls.PARAM_CACHE_TYPES),
-                }
-            }
-        )
-        repo_kwargs: dict = {
-            "rev": config.get(cls.PARAM_REV),
-            "subrepos": config.get(cls.PARAM_SUBREPOS, False),
-            "uninitialized": True,
-        }
-        factory: Optional["RepoFactory"] = None
-        if url is None:
-            repo_kwargs["config"] = cache_config
-        else:
-            repo_kwargs["cache_dir"] = cache_dir
-            factory = erepo_factory(url, cache_config)
-
-        with _open(
-            url if url else root,
-            **repo_kwargs,
-        ) as repo:
-            return repo, factory
-
-    def _get_repo(self, path: str) -> "Repo":
-        """Returns repo that the path falls in, using prefix.
-
-        If the path is already tracked/collected, it just returns the repo.
-
-        Otherwise, it collects the repos that might be in the path's parents
-        and then returns the appropriate one.
-        """
-        if not self.repo.fs.path.isin_or_eq(path, self.repo.root_dir):
-            # outside of repo
-            return self.repo
-
-        key = self._get_key(path)
-        repo = self._subrepos_trie.get(key)
-        if repo:
-            return repo
-
-        prefix_key, repo = self._subrepos_trie.longest_prefix(key)
-        prefix = self.repo.fs.path.join(
-            self.repo.root_dir,
-            *prefix_key,  # pylint: disable=not-an-iterable
-        )
-
-        parents = (parent for parent in self.repo.fs.path.parents(path))
-        dirs = [path] + list(takewhile(lambda p: p != prefix, parents))
-        dirs.reverse()
-        self._update(dirs, starting_repo=repo)
-        return self._subrepos_trie.get(key) or self.repo
-
-    @wrap_with(threading.Lock())
-    def _update(self, dirs, starting_repo):
-        """Checks for subrepo in directories and updates them."""
-        repo = starting_repo
-        for d in dirs:
-            key = self._get_key(d)
-            if self._is_dvc_repo(d):
-                repo = self.repo_factory(
-                    d,
-                    fs=self.repo.fs,
-                    repo_factory=self.repo_factory,
-                )
-                self._dvcfss[key] = DvcFileSystem(repo=repo)
-            self._subrepos_trie[key] = repo
-
-    def _is_dvc_repo(self, dir_path):
-        """Check if the directory is a dvc repo."""
-        if not self._traverse_subrepos:
-            return False
-
-        from dvc.repo import Repo
-
-        repo_path = os.path.join(dir_path, Repo.DVC_DIR)
-        return self.repo.fs.isdir(repo_path)
-
-    def _get_fs_pair(
-        self, path
-    ) -> Tuple[
-        Optional[FileSystem],
-        Optional[str],
-        Optional[DvcFileSystem],
-        Optional[str],
-    ]:
-        """
-        Returns a pair of fss based on repo the path falls in, using prefix.
-        """
-        from dvc.utils import as_posix
-
-        if os.path.isabs(path):
-            if self.repo.fs.path.isin_or_eq(path, self.repo.root_dir):
-                path = self.repo.fs.path.relpath(path, self.repo.root_dir)
-            else:
-                return None, None, self.repo.dvcfs, path
-
-        path = as_posix(path)
-
-        parts = self.path.parts(path)
-        if parts and parts[0] == os.curdir:
-            parts = parts[1:]
-
-        fs_path = self.repo.fs.path.join(self.repo.root_dir, *parts)
-        repo = self._get_repo(fs_path)
-        fs = repo.fs
-
-        repo_parts = fs.path.relparts(repo.root_dir, self.repo.root_dir)
-        if repo_parts[0] == os.curdir:
-            repo_parts = repo_parts[1:]
-
-        dvc_parts = parts[len(repo_parts) :]
-        if dvc_parts and dvc_parts[0] == os.curdir:
-            dvc_parts = dvc_parts[1:]
-
-        key = self._get_key(repo.root_dir)
-        dvc_fs = self._dvcfss.get(key)
-        if dvc_fs:
-            dvc_path = dvc_fs.path.join(*dvc_parts) if dvc_parts else ""
-        else:
-            dvc_path = None
-
-        return fs, fs_path, dvc_fs, dvc_path
-
-    def open(
-        self, path, mode="r", encoding="utf-8", **kwargs
-    ):  # pylint: disable=arguments-renamed, arguments-differ
-        if "b" in mode:
-            encoding = None
-
-        fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
-        try:
-            return fs.open(fs_path, mode=mode, encoding=encoding)
-        except FileNotFoundError:
-            if not dvc_fs:
-                raise
-
-        return dvc_fs.open(dvc_path, mode=mode, encoding=encoding, **kwargs)
-
-    def isdvc(self, path, **kwargs):
-        _, _, dvc_fs, dvc_path = self._get_fs_pair(path)
-        return dvc_fs is not None and dvc_fs.isdvc(dvc_path, **kwargs)
-
-    def ls(self, path, detail=True, **kwargs):
-        fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
-
-        repo = dvc_fs.repo if dvc_fs else self.repo
-        dvcignore = repo.dvcignore
-        ignore_subrepos = kwargs.get("ignore_subrepos", True)
-
-        names = set()
-        if dvc_fs:
-            with suppress(FileNotFoundError):
-                for entry in dvc_fs.ls(dvc_path, detail=False):
-                    names.add(dvc_fs.path.name(entry))
-
-        if fs:
-            try:
-                for entry in dvcignore.ls(
-                    fs, fs_path, detail=False, ignore_subrepos=ignore_subrepos
-                ):
-                    names.add(fs.path.name(entry))
-            except (FileNotFoundError, NotADirectoryError):
-                pass
-
-        dvcfiles = kwargs.get("dvcfiles", False)
-
-        def _func(fname):
-            from dvc.dvcfile import is_valid_filename
-            from dvc.ignore import DvcIgnore
-
-            if dvcfiles:
-                return True
-
-            return not (
-                is_valid_filename(fname) or fname == DvcIgnore.DVCIGNORE_FILE
-            )
-
-        names = filter(_func, names)
-
-        infos = []
-        paths = []
-        for name in names:
-            entry_path = self.path.join(path, name)
-            try:
-                info = self.info(entry_path, ignore_subrepos=ignore_subrepos)
-            except FileNotFoundError:
-                continue
-            infos.append(info)
-            paths.append(entry_path)
-
-        if not detail:
-            return paths
-
-        return infos
-
-    def get_file(self, rpath, lpath, callback=DEFAULT_CALLBACK, **kwargs):
-        fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(rpath)
-
-        if fs:
-            try:
-                fs.get_file(  # pylint: disable=protected-access
-                    fs_path, lpath, callback=callback, **kwargs
-                )
-                return
-            except FileNotFoundError:
-                if not dvc_fs:
-                    raise
-
-        dvc_fs.get_file(  # pylint: disable=protected-access
-            dvc_path, lpath, callback=callback, **kwargs
-        )
-
-    def info(self, path, **kwargs):
-        fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
-
-        repo = dvc_fs.repo if dvc_fs else self.repo
-        dvcignore = repo.dvcignore
-        ignore_subrepos = kwargs.get("ignore_subrepos", True)
-
-        dvc_info = None
-        if dvc_fs:
-            try:
-                dvc_info = dvc_fs.info(dvc_path)
-            except FileNotFoundError:
-                pass
-
-        fs_info = None
-        if fs:
-            try:
-                fs_info = fs.info(fs_path)
-                if dvcignore.is_ignored(
-                    fs, fs_path, ignore_subrepos=ignore_subrepos
-                ):
-                    fs_info = None
-            except (FileNotFoundError, NotADirectoryError):
-                if not dvc_info:
-                    raise
-
-        # NOTE: if some parent in fs_path turns out to be a file, it means
-        # that the whole repofs branch doesn't exist.
-        if fs and not fs_info and dvc_info:
-            for parent in fs.path.parents(fs_path):
-                try:
-                    if fs.info(parent)["type"] != "directory":
-                        dvc_info = None
-                        break
-                except FileNotFoundError:
-                    continue
-
-        if not dvc_info and not fs_info:
-            raise FileNotFoundError
-
-        info = _merge_info(dvc_fs.repo, fs_info, dvc_info)
-        info["name"] = path
-        return info
-
-    def checksum(self, path):
-        fs, fs_path, dvc_fs, dvc_path = self._get_fs_pair(path)
-
-        try:
-            return fs.checksum(fs_path)
-        except FileNotFoundError:
-            return dvc_fs.checksum(dvc_path)
-
-
-class RepoFileSystem(FileSystem):
-    scheme = "local"
-    PARAM_CHECKSUM = "md5"
-
-    def _prepare_credentials(self, **config):
-        return config
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        return _RepoFileSystem(**self.fs_args)
-
-    def isdvc(self, path, **kwargs):
-        return self.fs.isdvc(path, **kwargs)
-
-    @property
-    def repo(self):
-        return self.fs.repo
-
-    @property
-    def repo_url(self):
-        return self.fs.repo_url
-
-    @property
-    def config(self):
-        return self.fs.config

dvc/fs/s3.py

@@ -1,160 +0,0 @@
-import os
-import threading
-from collections import defaultdict
-
-from funcy import cached_property, wrap_prop
-
-from dvc.scheme import Schemes
-
-from .base import ObjectFileSystem
-
-_AWS_CONFIG_PATH = os.path.join(os.path.expanduser("~"), ".aws", "config")
-
-
-# pylint:disable=abstract-method
-class S3FileSystem(ObjectFileSystem):
-    scheme = Schemes.S3
-    REQUIRES = {"s3fs": "s3fs", "boto3": "boto3"}
-    PARAM_CHECKSUM = "etag"
-
-    _GRANTS = {
-        "grant_full_control": "GrantFullControl",
-        "grant_read": "GrantRead",
-        "grant_read_acp": "GrantReadACP",
-        "grant_write_acp": "GrantWriteACP",
-    }
-
-    _TRANSFER_CONFIG_ALIASES = {
-        "max_queue_size": "max_io_queue",
-        "max_concurrent_requests": "max_concurrency",
-        "multipart_threshold": "multipart_threshold",
-        "multipart_chunksize": "multipart_chunksize",
-    }
-
-    def _split_s3_config(self, s3_config):
-        """Splits the general s3 config into 2 different config
-        objects, one for transfer.TransferConfig and other is the
-        general session config"""
-
-        from boto3.s3.transfer import TransferConfig
-
-        from dvc.utils import conversions
-
-        config, transfer_config = {}, {}
-        for key, value in s3_config.items():
-            if key in self._TRANSFER_CONFIG_ALIASES:
-                if key in {"multipart_chunksize", "multipart_threshold"}:
-                    # cast human readable sizes (like 24MiB) to integers
-                    value = conversions.human_readable_to_bytes(value)
-                else:
-                    value = int(value)
-                transfer_config[self._TRANSFER_CONFIG_ALIASES[key]] = value
-            else:
-                config[key] = value
-
-        # pylint: disable=attribute-defined-outside-init
-        self._transfer_config = TransferConfig(**transfer_config)
-        return config
-
-    def _load_aws_config_file(self, profile):
-        from botocore.configloader import load_config
-
-        # pylint: disable=attribute-defined-outside-init
-        self._transfer_config = None
-        config_path = os.environ.get("AWS_CONFIG_FILE", _AWS_CONFIG_PATH)
-        if not os.path.exists(config_path):
-            return {}
-
-        config = load_config(config_path)
-        profile_config = config["profiles"].get(profile or "default")
-        if not profile_config:
-            return {}
-
-        s3_config = profile_config.get("s3", {})
-        return self._split_s3_config(s3_config)
-
-    def _prepare_credentials(self, **config):
-        from dvc.config import ConfigError
-        from dvc.utils.flatten import flatten, unflatten
-
-        login_info = defaultdict(dict)
-
-        # credentials
-        login_info["key"] = config.get("access_key_id")
-        login_info["secret"] = config.get("secret_access_key")
-        login_info["token"] = config.get("session_token")
-
-        # session configuration
-        login_info["profile"] = config.get("profile")
-        login_info["use_ssl"] = config.get("use_ssl", True)
-
-        # extra client configuration
-        client = login_info["client_kwargs"]
-        client["region_name"] = config.get("region")
-        client["endpoint_url"] = config.get("endpointurl")
-        client["verify"] = config.get("ssl_verify")
-
-        # timeout configuration
-        config_kwargs = login_info["config_kwargs"]
-        config_kwargs["read_timeout"] = config.get("read_timeout")
-        config_kwargs["connect_timeout"] = config.get("connect_timeout")
-
-        # encryptions
-        additional = login_info["s3_additional_kwargs"]
-        additional["ServerSideEncryption"] = config.get("sse")
-        additional["SSEKMSKeyId"] = config.get("sse_kms_key_id")
-        additional["ACL"] = config.get("acl")
-        for grant_option, grant_key in self._GRANTS.items():
-            if config.get(grant_option):
-                if additional["ACL"]:
-                    raise ConfigError(
-                        "`acl` and `grant_*` AWS S3 config options "
-                        "are mutually exclusive"
-                    )
-                additional[grant_key] = config[grant_option]
-
-        # config kwargs
-        session_config = login_info["config_kwargs"]
-        session_config["s3"] = self._load_aws_config_file(
-            login_info["profile"]
-        )
-
-        shared_creds = config.get("credentialpath")
-        if shared_creds:
-            os.environ.setdefault("AWS_SHARED_CREDENTIALS_FILE", shared_creds)
-
-        if (
-            client["region_name"] is None
-            and session_config["s3"].get("region_name") is None
-            and os.getenv("AWS_REGION") is None
-        ):
-            # Enable bucket region caching
-            login_info["cache_regions"] = config.get("cache_regions", True)
-
-        config_path = config.get("configpath")
-        if config_path:
-            os.environ.setdefault("AWS_CONFIG_FILE", config_path)
-
-        return unflatten(
-            {
-                key: value
-                for key, value in flatten(login_info).items()
-                if value is not None
-            }
-        )
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        from s3fs import S3FileSystem as _S3FileSystem
-
-        return _S3FileSystem(**self.fs_args)
-
-    @classmethod
-    def _strip_protocol(cls, path: str) -> str:
-        from fsspec.utils import infer_storage_options
-
-        return infer_storage_options(path)["path"]
-
-    def unstrip_protocol(self, path):
-        return "s3://" + path.lstrip("/")

dvc/fs/ssh.py

@@ -1,140 +0,0 @@
-import getpass
-import os.path
-import threading
-
-from funcy import cached_property, memoize, silent, wrap_prop, wrap_with
-
-from dvc import prompt
-from dvc.scheme import Schemes
-from dvc.utils.fs import as_atomic
-
-from ._callback import DEFAULT_CALLBACK
-from .base import FileSystem
-
-DEFAULT_PORT = 22
-
-
-@wrap_with(threading.Lock())
-@memoize
-def ask_password(host, user, port):
-    return prompt.password(
-        "Enter a private key passphrase or a password for "
-        "host '{host}' port '{port}' user '{user}'".format(
-            host=host, port=port, user=user
-        )
-    )
-
-
-# pylint:disable=abstract-method
-class SSHFileSystem(FileSystem):
-    scheme = Schemes.SSH
-    REQUIRES = {"sshfs": "sshfs"}
-    PARAM_CHECKSUM = "md5"
-
-    @classmethod
-    def _strip_protocol(cls, path: str) -> str:
-        from fsspec.utils import infer_storage_options
-
-        return infer_storage_options(path)["path"]
-
-    def unstrip_protocol(self, path: str) -> str:
-        host = self.fs_args["host"]
-        port = self.fs_args["port"]
-        path = path.lstrip("/")
-        return f"ssh://{host}:{port}/{path}"
-
-    def _prepare_credentials(self, **config):
-        self.CAN_TRAVERSE = True
-        from sshfs.config import parse_config
-
-        login_info = {}
-
-        try:
-            user_ssh_config = parse_config(host=config["host"])
-        except FileNotFoundError:
-            user_ssh_config = {}
-
-        login_info["host"] = user_ssh_config.get("Hostname", config["host"])
-
-        login_info["username"] = (
-            config.get("user")
-            or config.get("username")
-            or user_ssh_config.get("User")
-            or getpass.getuser()
-        )
-        login_info["port"] = (
-            config.get("port")
-            or silent(int)(user_ssh_config.get("Port"))
-            or DEFAULT_PORT
-        )
-
-        if config.get("ask_password") and config.get("password") is None:
-            config["password"] = ask_password(
-                login_info["host"], login_info["username"], login_info["port"]
-            )
-
-        login_info["password"] = config.get("password")
-        login_info["passphrase"] = config.get("password")
-
-        raw_keys = []
-        if config.get("keyfile"):
-            raw_keys.append(config.get("keyfile"))
-        elif user_ssh_config.get("IdentityFile"):
-            raw_keys.extend(user_ssh_config.get("IdentityFile"))
-
-        if raw_keys:
-            login_info["client_keys"] = [
-                os.path.expanduser(key) for key in raw_keys
-            ]
-
-        login_info["timeout"] = config.get("timeout", 1800)
-
-        # These two settings fine tune the asyncssh to use the
-        # fastest encryption algorithm and disable compression
-        # altogether (since it blocking, it is slowing down
-        # the transfers in a considerable rate, and even for
-        # compressible data it is making it extremely slow).
-        # See: https://github.com/ronf/asyncssh/issues/374
-        login_info["encryption_algs"] = [
-            "[email protected]",
-            "aes256-ctr",
-            "aes192-ctr",
-            "aes128-ctr",
-        ]
-        login_info["compression_algs"] = None
-
-        login_info["gss_auth"] = config.get("gss_auth", False)
-        login_info["agent_forwarding"] = config.get("agent_forwarding", True)
-        login_info["proxy_command"] = user_ssh_config.get("ProxyCommand")
-
-        # We are going to automatically add stuff to known_hosts
-        # something like paramiko's AutoAddPolicy()
-        login_info["known_hosts"] = None
-        return login_info
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        from sshfs import SSHFileSystem as _SSHFileSystem
-
-        return _SSHFileSystem(**self.fs_args)
-
-    # Ensure that if an interrupt happens during the transfer, we don't
-    # pollute the cache.
-
-    def upload_fobj(self, fobj, to_info, **kwargs):
-        with as_atomic(self, to_info) as tmp_file:
-            super().upload_fobj(fobj, tmp_file, **kwargs)
-
-    def put_file(
-        self,
-        from_file,
-        to_info,
-        callback=DEFAULT_CALLBACK,
-        size=None,
-        **kwargs,
-    ):
-        with as_atomic(self, to_info) as tmp_file:
-            super().put_file(
-                from_file, tmp_file, callback=callback, size=size, **kwargs
-            )

dvc/fs/utils.py

@@ -1,169 +0,0 @@
-import errno
-import logging
-import os
-from typing import TYPE_CHECKING, List, Optional
-
-from .base import RemoteActionNotImplemented
-from .local import LocalFileSystem
-
-if TYPE_CHECKING:
-    from .base import FileSystem
-    from .types import AnyPath
-
-logger = logging.getLogger(__name__)
-
-
-def _link(
-    link: "AnyPath",
-    from_fs: "FileSystem",
-    from_path: "AnyPath",
-    to_fs: "FileSystem",
-    to_path: "AnyPath",
-) -> None:
-    if not isinstance(from_fs, type(to_fs)):
-        raise OSError(errno.EXDEV, "can't link across filesystems")
-
-    try:
-        func = getattr(to_fs, link)
-        func(from_path, to_path)
-    except (OSError, AttributeError, RemoteActionNotImplemented) as exc:
-        # NOTE: there are so many potential error codes that a link call can
-        # throw, that is safer to just catch them all and try another link.
-        raise OSError(
-            errno.ENOTSUP, f"'{link}' is not supported by {type(from_fs)}"
-        ) from exc
-
-
-def _copy(
-    from_fs: "FileSystem",
-    from_path: "AnyPath",
-    to_fs: "FileSystem",
-    to_path: "AnyPath",
-) -> None:
-    if isinstance(from_fs, LocalFileSystem):
-        return to_fs.upload(from_path, to_path)
-
-    if isinstance(to_fs, LocalFileSystem):
-        return from_fs.download_file(from_path, to_path)
-
-    with from_fs.open(from_path, mode="rb") as fobj:
-        size = from_fs.size(from_path)
-        return to_fs.upload(fobj, to_path, size=size)
-
-
-def _try_links(
-    links: List["AnyPath"],
-    from_fs: "FileSystem",
-    from_path: "AnyPath",
-    to_fs: "FileSystem",
-    to_path: "AnyPath",
-) -> None:
-    error = None
-    while links:
-        link = links[0]
-
-        if link == "copy":
-            return _copy(from_fs, from_path, to_fs, to_path)
-
-        try:
-            return _link(link, from_fs, from_path, to_fs, to_path)
-        except OSError as exc:
-            if exc.errno not in [errno.ENOTSUP, errno.EXDEV]:
-                raise
-            error = exc
-
-        del links[0]
-
-    raise OSError(
-        errno.ENOTSUP, "no more link types left to try out"
-    ) from error
-
-
-def transfer(
-    from_fs: "FileSystem",
-    from_path: "AnyPath",
-    to_fs: "FileSystem",
-    to_path: "AnyPath",
-    hardlink: bool = False,
-    links: Optional[List["AnyPath"]] = None,
-) -> None:
-    try:
-        assert not (hardlink and links)
-        if hardlink:
-            links = links or ["reflink", "hardlink", "copy"]
-        else:
-            links = links or ["reflink", "copy"]
-        _try_links(links, from_fs, from_path, to_fs, to_path)
-    except OSError as exc:
-        # If the target file already exists, we are going to simply
-        # ignore the exception (#4992).
-        #
-        # On Windows, it is not always guaranteed that you'll get
-        # FileExistsError (it might be PermissionError or a bare OSError)
-        # but all of those exceptions raised from the original
-        # FileExistsError so we have a separate check for that.
-        if isinstance(exc, FileExistsError) or (
-            os.name == "nt"
-            and exc.__context__
-            and isinstance(exc.__context__, FileExistsError)
-        ):
-            logger.debug("'%s' file already exists, skipping", to_path)
-            return None
-
-        raise
-
-
-def _test_link(
-    link: "AnyPath",
-    from_fs: "FileSystem",
-    from_file: "AnyPath",
-    to_fs: "FileSystem",
-    to_file: "AnyPath",
-) -> bool:
-    try:
-        _try_links([link], from_fs, from_file, to_fs, to_file)
-    except OSError:
-        logger.debug("", exc_info=True)
-        return False
-
-    try:
-        _is_link_func = getattr(to_fs, f"is_{link}")
-        return _is_link_func(to_file)
-    except AttributeError:
-        pass
-
-    return True
-
-
-def test_links(
-    links: List["AnyPath"],
-    from_fs: "FileSystem",
-    from_path: "AnyPath",
-    to_fs: "FileSystem",
-    to_path: "AnyPath",
-) -> List["AnyPath"]:
-    from dvc.utils import tmp_fname
-
-    from_file = from_fs.path.join(from_path, tmp_fname())
-    to_file = to_fs.path.join(
-        to_fs.path.parent(to_path),
-        tmp_fname(),
-    )
-
-    from_fs.makedirs(from_fs.path.parent(from_file))
-    with from_fs.open(from_file, "wb") as fobj:
-        fobj.write(b"test")
-    to_fs.makedirs(to_fs.path.parent(to_file))
-
-    ret = []
-    try:
-        for link in links:
-            try:
-                if _test_link(link, from_fs, from_file, to_fs, to_file):
-                    ret.append(link)
-            finally:
-                to_fs.remove(to_file)
-    finally:
-        from_fs.remove(from_file)
-
-    return ret

dvc/fs/webdav.py

@@ -1,92 +0,0 @@
-import logging
-import threading
-
-from funcy import cached_property, wrap_prop
-
-from dvc.scheme import Schemes
-
-from .base import FileSystem
-from .http import ask_password
-
-logger = logging.getLogger(__name__)
-
-
-class WebDAVFileSystem(FileSystem):  # pylint:disable=abstract-method
-    scheme = Schemes.WEBDAV
-    root_marker = ""
-    REQUIRES = {"webdav4": "webdav4"}
-    PARAM_CHECKSUM = "etag"
-
-    def __init__(self, **config):
-        super().__init__(**config)
-
-        cert_path = config.get("cert_path", None)
-        key_path = config.get("key_path", None)
-        self.prefix = config.get("prefix", "")
-        self.fs_args.update(
-            {
-                "base_url": config["url"],
-                "cert": cert_path if not key_path else (cert_path, key_path),
-                "verify": config.get("ssl_verify", True),
-                "timeout": config.get("timeout", 30),
-            }
-        )
-
-    def unstrip_protocol(self, path: str) -> str:
-        return self.fs_args["base_url"] + "/" + path
-
-    @staticmethod
-    def _get_kwargs_from_urls(urlpath):
-        from urllib.parse import urlparse, urlunparse
-
-        parsed = urlparse(urlpath)
-        scheme = parsed.scheme.replace("webdav", "http")
-        return {
-            "prefix": parsed.path.rstrip("/"),
-            "host": urlunparse((scheme, parsed.netloc, "", None, None, None)),
-            "url": urlunparse(
-                (
-                    scheme,
-                    parsed.netloc,
-                    parsed.path.rstrip("/"),
-                    None,
-                    None,
-                    None,
-                )
-            ),
-            "user": parsed.username,
-        }
-
-    def _prepare_credentials(self, **config):
-        user = config.get("user", None)
-        password = config.get("password", None)
-
-        headers = {}
-        token = config.get("token")
-        auth = None
-        if token:
-            headers.update({"Authorization": f"Bearer {token}"})
-        elif user:
-            if not password and config.get("ask_password"):
-                password = ask_password(config["host"], user)
-            auth = (user, password)
-
-        return {"headers": headers, "auth": auth}
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        from webdav4.fsspec import WebdavFileSystem
-
-        return WebdavFileSystem(**self.fs_args)
-
-    def upload_fobj(self, fobj, to_info, **kwargs):
-        size = kwargs.get("size")
-        # using upload_fileobj to directly upload fileobj rather than buffering
-        # and using overwrite=True to avoid check for an extra exists call,
-        # as caller should ensure that the file does not exist beforehand.
-        return self.fs.upload_fileobj(fobj, to_info, overwrite=True, size=size)
-
-
-class WebDAVSFileSystem(WebDAVFileSystem):  # pylint:disable=abstract-method
-    scheme = Schemes.WEBDAVS

dvc/fs/webhdfs.py

@@ -1,56 +0,0 @@
-import threading
-
-from funcy import cached_property, wrap_prop
-
-from dvc.scheme import Schemes
-
-# pylint:disable=abstract-method
-from .base import FileSystem
-
-
-class WebHDFSFileSystem(FileSystem):
-    scheme = Schemes.WEBHDFS
-    REQUIRES = {"fsspec": "fsspec"}
-    PARAM_CHECKSUM = "checksum"
-
-    @classmethod
-    def _strip_protocol(cls, path: str) -> str:
-        from fsspec.utils import infer_storage_options
-
-        return infer_storage_options(path)["path"]
-
-    def unstrip_protocol(self, path: str) -> str:
-        host = self.fs_args["host"]
-        port = self.fs_args["port"]
-        path = path.lstrip("/")
-        return f"webhdfs://{host}:{port}/{path}"
-
-    @staticmethod
-    def _get_kwargs_from_urls(urlpath):
-        from fsspec.implementations.webhdfs import WebHDFS
-
-        return (
-            WebHDFS._get_kwargs_from_urls(  # pylint:disable=protected-access
-                urlpath
-            )
-        )
-
-    def _prepare_credentials(self, **config):
-        self._ssl_verify = config.pop("ssl_verify", True)
-        principal = config.pop("kerberos_principal", None)
-        if principal:
-            config["kerb_kwargs"] = {"principal": principal}
-        return config
-
-    @wrap_prop(threading.Lock())
-    @cached_property
-    def fs(self):
-        from fsspec.implementations.webhdfs import WebHDFS
-
-        fs = WebHDFS(**self.fs_args)
-        fs.session.verify = self._ssl_verify
-        return fs
-
-    def checksum(self, path):
-        ukey = self.fs.ukey(path)
-        return ukey["bytes"]

dvc/hash_info.py

@@ -1,43 +0,0 @@
-from collections import OrderedDict
-from dataclasses import dataclass, field
-from typing import Optional
-
-HASH_DIR_SUFFIX = ".dir"
-
-
-@dataclass
-class HashInfo:
-    name: Optional[str]
-    value: Optional[str]
-    obj_name: Optional[str] = field(default=None, compare=False)
-
-    def __bool__(self):
-        return bool(self.value)
-
-    def __str__(self):
-        return f"{self.name}: {self.value}"
-
-    def __hash__(self):
-        return hash((self.name, self.value))
-
-    @classmethod
-    def from_dict(cls, d):
-        if not d:
-            return cls(None, None)
-
-        ((name, value),) = d.items()
-        return cls(name, value)
-
-    def to_dict(self):
-        ret = OrderedDict()
-        if not self:
-            return ret
-
-        ret[self.name] = self.value
-        return ret
-
-    @property
-    def isdir(self):
-        if not self:
-            return False
-        return self.value.endswith(HASH_DIR_SUFFIX)

dvc/ignore.py

@@ -8,10 +8,8 @@ from pathspec.patterns import GitWildMatchPattern
 from pathspec.util import normalize_file
 from pygtrie import Trie
 
-from dvc.fs.base import AnyFSPath, FileSystem
-from dvc.fs.local import localfs
+from dvc.fs import AnyFSPath, FileSystem, Schemes, localfs
 from dvc.pathspec_math import PatternInfo, merge_patterns
-from dvc.scheme import Schemes
 from dvc.types import List, Optional
 
 logger = logging.getLogger(__name__)
@@ -35,7 +33,6 @@ class DvcIgnorePatterns(DvcIgnore):
         self.sep = sep
         self.pattern_list = pattern_list
         self.dirname = dirname
-        self.prefix = self.dirname + self.sep
 
         self.regex_pattern_list = [
             GitWildMatchPattern.pattern_to_regex(pattern_info.patterns)
@@ -74,10 +71,13 @@ class DvcIgnorePatterns(DvcIgnore):
     def _get_normalize_path(self, dirname, basename):
         # NOTE: `relpath` is too slow, so we have to assume that both
         # `dirname` and `self.dirname` are relative or absolute together.
+
+        prefix = self.dirname.rstrip(self.sep) + self.sep
+
         if dirname == self.dirname:
             path = basename
-        elif dirname.startswith(self.prefix):
-            rel = dirname[len(self.prefix) :]
+        elif dirname.startswith(prefix):
+            rel = dirname[len(prefix) :]
             # NOTE: `os.path.join` is ~x5.5 slower
             path = f"{rel}{self.sep}{basename}"
         else:
@@ -210,6 +210,7 @@ class DvcIgnoreFilter:
             new_pattern = DvcIgnorePatterns.from_file(path, self.fs, name)
             if old_pattern:
                 plist, prefix = merge_patterns(
+                    self.fs.path.flavour,
                     old_pattern.pattern_list,
                     old_pattern.dirname,
                     new_pattern.pattern_list,
@@ -259,6 +260,7 @@ class DvcIgnoreFilter:
         old_pattern = ignore_trie.longest_prefix(key).value
         if old_pattern:
             plist, prefix = merge_patterns(
+                self.fs.path.flavour,
                 old_pattern.pattern_list,
                 old_pattern.dirname,
                 new_pattern.pattern_list,
@@ -299,7 +301,7 @@ class DvcIgnoreFilter:
 
     def walk(self, fs: FileSystem, path: AnyFSPath, **kwargs):
         ignore_subrepos = kwargs.pop("ignore_subrepos", True)
-        if fs.scheme == Schemes.LOCAL:
+        if fs.protocol == Schemes.LOCAL:
             for root, dirs, files in fs.walk(path, **kwargs):
                 dirs[:], files[:] = self(
                     root, dirs, files, ignore_subrepos=ignore_subrepos
@@ -309,7 +311,7 @@ class DvcIgnoreFilter:
             yield from fs.walk(path, **kwargs)
 
     def find(self, fs: FileSystem, path: AnyFSPath, **kwargs):
-        if fs.scheme == Schemes.LOCAL:
+        if fs.protocol == Schemes.LOCAL:
             for root, _, files in self.walk(fs, path, **kwargs):
                 for file in files:
                     # NOTE: os.path.join is ~5.5 times slower
@@ -402,7 +404,7 @@ class DvcIgnoreFilter:
     ) -> bool:
         # NOTE: can't use self.check_ignore(path).match for now, see
         # https://github.com/iterative/dvc/issues/4555
-        if fs.scheme != Schemes.LOCAL:
+        if fs.protocol != Schemes.LOCAL:
             return False
         if fs.isfile(path):
             return self.is_ignored_file(path, ignore_subrepos)

dvc/info.py

@@ -1,3 +1,4 @@
+import importlib.metadata as importlib_metadata
 import itertools
 import os
 import pathlib
@@ -7,19 +8,12 @@ import psutil
 
 from dvc import __version__
 from dvc.exceptions import NotDvcRepoError
-from dvc.fs import FS_MAP, get_fs_cls, get_fs_config
-from dvc.fs.utils import test_links
+from dvc.fs import Schemes, generic, get_fs_cls, get_fs_config, registry
 from dvc.repo import Repo
 from dvc.scm import SCMError
 from dvc.utils import error_link
 from dvc.utils.pkg import PKG
 
-try:
-    import importlib.metadata as importlib_metadata
-except ImportError:  # < 3.8
-    import importlib_metadata  # type: ignore[no-redef]
-
-
 package = "" if PKG is None else f"({PKG})"
 
 
@@ -38,9 +32,9 @@ def get_dvc_info():
             # can't auto-create it, as it might cause issues if the user
             # later decides to enable shared cache mode with
             # `dvc config cache.shared group`.
-            if os.path.exists(repo.odb.local.cache_dir):
+            if os.path.exists(repo.odb.local.path):
                 info.append(f"Cache types: {_get_linktype_support_info(repo)}")
-                fs_type = get_fs_type(repo.odb.local.cache_dir)
+                fs_type = get_fs_type(repo.odb.local.path)
                 info.append(f"Cache directory: {fs_type}")
             else:
                 info.append("Cache types: " + error_link("no-dvc-cache"))
@@ -64,7 +58,7 @@ def _get_caches(cache):
     caches = (
         cache_type
         for cache_type, cache_instance in cache.by_scheme()
-        if cache_instance
+        if cache_instance and cache_type != "repo"
     )
 
     # Caches will be always non-empty including the local cache
@@ -73,7 +67,7 @@ def _get_caches(cache):
 
 def _get_remotes(config):
     schemes = (
-        get_fs_cls(get_fs_config(None, config, name=remote)).scheme
+        get_fs_cls(get_fs_config(None, config, name=remote)).protocol
         for remote in config["remote"]
     )
 
@@ -83,10 +77,10 @@ def _get_remotes(config):
 def _get_linktype_support_info(repo):
     odb = repo.odb.local
 
-    links = test_links(
+    links = generic.test_links(
         ["reflink", "hardlink", "symlink"],
         odb.fs,
-        odb.fs_path,
+        odb.path,
         repo.fs,
         repo.root_dir,
     )
@@ -96,7 +90,15 @@ def _get_linktype_support_info(repo):
 
 def _get_supported_remotes():
     supported_remotes = []
-    for scheme, fs_cls in FS_MAP.items():
+    for scheme in registry:
+        if scheme in [Schemes.LOCAL, Schemes.MEMORY, "dvc", "git"]:
+            continue
+
+        try:
+            fs_cls = registry[scheme]
+        except ImportError:
+            continue
+
         if not fs_cls.get_missing_deps():
             dependencies = []
             for requirement in fs_cls.REQUIRES:

dvc/istextfile.py

@@ -1,34 +0,0 @@
-"""Use heuristics to guess if it is a text file or a binary file."""
-
-# Based on https://eli.thegreenplace.net/2011/10/19/
-# perls-guess-if-file-is-text-or-binary-implemented-in-python
-
-
-TEXT_CHARS = bytes(range(32, 127)) + b"\n\r\t\f\b"
-DEFAULT_CHUNK_SIZE = 512
-
-
-def istextblock(block):
-    if not block:
-        # An empty file is considered a valid text file
-        return True
-
-    if b"\x00" in block:
-        # Files with null bytes are binary
-        return False
-
-    # Use translate's 'deletechars' argument to efficiently remove all
-    # occurrences of TEXT_CHARS from the block
-    nontext = block.translate(None, TEXT_CHARS)
-    return float(len(nontext)) / len(block) <= 0.30
-
-
-def istextfile(fname, fs, blocksize=DEFAULT_CHUNK_SIZE):
-    """Uses heuristics to guess whether the given file is text or binary,
-    by reading a single block of bytes from the file.
-    If more than 30% of the chars in the block are non-text, or there
-    are NUL ('\x00') bytes in the block, assume this is a binary file.
-    """
-    with fs.open(fname, "rb") as fobj:
-        block = fobj.read(blocksize)
-    return istextblock(block)

dvc/lock.py

@@ -70,6 +70,8 @@ class LockNoop(LockBase):
         self._lock = True
 
     def unlock(self):
+        if not self.is_locked:
+            raise DvcException("Unlock called on an unlocked lock")
         self._lock = False
 
     @property
@@ -93,6 +95,7 @@ class Lock(LockBase):
         super().__init__(lockfile)
         self._friendly = friendly
         self._lock = None
+        self._lock_failed = False
 
     @property
     def files(self):
@@ -100,6 +103,7 @@ class Lock(LockBase):
 
     def _do_lock(self):
         try:
+            self._lock_failed = False
             with Tqdm(
                 bar_format="{desc}",
                 disable=not self._friendly,
@@ -111,6 +115,7 @@ class Lock(LockBase):
             ):
                 self._lock = zc.lockfile.LockFile(self._lockfile)
         except zc.lockfile.LockError:
+            self._lock_failed = True
             raise LockError(FAILED_TO_LOCK_MESSAGE)
 
     def lock(self):
@@ -120,6 +125,12 @@ class Lock(LockBase):
         lock_retry()
 
     def unlock(self):
+        if self._lock_failed:
+            assert self._lock is None
+            return
+
+        if not self.is_locked:
+            raise DvcException("Unlock called on an unlocked lock")
         self._lock.close()
         self._lock = None
 

dvc/logger.py

@@ -163,7 +163,7 @@ class LoggerHandler(logging.StreamHandler):
                 msg, file=self.stream, end=getattr(self, "terminator", "\n")
             )
             self.flush()
-        except RecursionError:
+        except (BrokenPipeError, RecursionError):
             raise
         except Exception:  # noqa, pylint: disable=broad-except
             self.handleError(record)
@@ -207,7 +207,12 @@ def disable_other_loggers():
             logger.disabled = True
 
 
-def setup(level=logging.INFO):
+def set_loggers_level(level: int = logging.INFO) -> None:
+    for name in ["dvc", "dvc_objects", "dvc_data"]:
+        logging.getLogger(name).setLevel(level)
+
+
+def setup(level: int = logging.INFO) -> None:
     colorama.init()
 
     if level >= logging.DEBUG:
@@ -269,7 +274,25 @@ def setup(level=logging.INFO):
                         "console_trace",
                         "console_errors",
                     ],
-                }
+                },
+                "dvc_objects": {
+                    "level": level,
+                    "handlers": [
+                        "console_info",
+                        "console_debug",
+                        "console_trace",
+                        "console_errors",
+                    ],
+                },
+                "dvc_data": {
+                    "level": level,
+                    "handlers": [
+                        "console_info",
+                        "console_debug",
+                        "console_trace",
+                        "console_errors",
+                    ],
+                },
             },
             "disable_existing_loggers": False,
         }

dvc/machine/backend/base.py

@@ -3,7 +3,8 @@ from contextlib import contextmanager
 from typing import TYPE_CHECKING, Iterator, Optional
 
 if TYPE_CHECKING:
-    from dvc.fs.ssh import SSHFileSystem
+    from dvc_ssh import SSHFileSystem
+
     from dvc.types import StrPath