Compare commits

...
65 Commits
Author SHA1 Message Date
George KussumotoandSaugat Pachhai 0e38ccc031 fix(pull): normalize dependency path (#10446)
Co-authored-by: Saugat Pachhai (सौगात) <[email protected]>
2024-06-03 16:43:43 +05:45
Ivan Shcheklein b69b368ffe fix(benchmarks): force older pygit2 for older DVC versions (#10447) 2024-06-03 15:56:43 +05:45
6cc80fdeaa fix fish shell support (#10445)
Co-authored-by: Alexey Stern [email protected]
Co-authored-by: Saugat Pachhai (सौगात) <[email protected]>
2024-06-03 15:45:21 +05:45
Ivan Shcheklein 8f8d49d023 fix(push/pull): don't fail on unsupported run cache (#10441) 2024-05-27 14:37:44 +05:45
Dave Berenbaum 83ec952da1 improve repro --pull --run-cache test 2024-05-23 14:26:42 -04:00
Dave Berenbaum c0d7e2a8de test repro --pull failure 2024-05-23 14:26:42 -04:00
Dave Berenbaum 9aad09164c update repro --pull description 2024-05-23 14:26:42 -04:00
Dave Berenbaum 1f7733fec1 refactor 2024-05-23 14:26:42 -04:00
Dave Berenbaum 856c902b65 add tests 2024-05-23 14:26:42 -04:00
Dave Berenbaum ebb459f9e2 refactor dvc repro --pull 2024-05-23 14:26:42 -04:00
Dave Berenbaum 2e0b024efe Apply suggestions from code review 2024-05-23 13:45:48 -04:00
Dave Berenbaum 135fa26515 Update dvc/repo/artifacts.py 2024-05-23 13:45:48 -04:00
Dave Berenbaum ef507efa04 artifacts: get studio info from local config 2024-05-23 13:45:48 -04:00
Dave Berenbaum 05db79e7cf add dvc artifacts get --show-url 2024-05-23 12:46:21 -04:00
Dave Berenbaum 0a813bbcbc push/pull run cache by default 2024-05-22 08:21:55 -04:00
skshetry e82356ba92 deps: bump scmrepo to >=3.3.4 (#10432)
Closes https://github.com/iterative/dvc/issues/10431.

Upstream PR: https://github.com/iterative/scmrepo/pull/359
2024-05-20 18:13:25 +05:45
Sigurd Spieckermann 5bfacec974 import/get: escape non-pattern LFS path parts (#10416) 2024-05-13 10:24:06 -07:00
pre-commit-ci[bot] 1848074ac9 [pre-commit.ci] pre-commit autoupdate (#10409)
* [pre-commit.ci] pre-commit autoupdate

updates:
- [github.com/astral-sh/ruff-pre-commit: v0.4.1 → v0.4.3](https://github.com/astral-sh/ruff-pre-commit/compare/v0.4.1...v0.4.3)

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2024-05-10 10:52:02 +00:00
Sigurd Spieckermann a7c0f6ef39 import/get: optimize LFS prefetching (#10414) 2024-05-03 08:58:53 -07:00
Thomas Roehr 0a8a9deceb Update s3 dependency to dvc-s3>=3.2.0 2024-04-26 11:01:03 -04:00
Thomas Roehr c5d98681c0 s3: align config name for anonymous login 2024-04-26 11:01:03 -04:00
Thomas Roehr 15e05e4b3e dvc: s3: allow configuring anonymous access 2024-04-26 11:01:03 -04:00
pre-commit-ci[bot] 7df8b0c09b [pre-commit.ci] pre-commit autoupdate (#10397)
updates:
- [github.com/astral-sh/ruff-pre-commit: v0.3.7 → v0.4.1](https://github.com/astral-sh/ruff-pre-commit/compare/v0.3.7...v0.4.1)

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2024-04-26 15:25:52 +05:45
Amirmohammad Sadat Shokouhi 2e1faf2979 add replication factor argument for hdfs in config schema (#10396) 2024-04-25 12:03:35 -07:00
skshetry cbaeff2ea1 deps: bump scmrepo minver to >=3.3.2,<4 (#10406)
Update pyproject.toml
2024-04-25 14:41:58 +05:45
dependabot[bot] 132e625ae8 build(deps): bump mypy from 1.9.0 to 1.10.0 (#10405)
Bumps [mypy](https://github.com/python/mypy) from 1.9.0 to 1.10.0.
- [Changelog](https://github.com/python/mypy/blob/master/CHANGELOG.md)
- [Commits](https://github.com/python/mypy/compare/1.9.0...v1.10.0)

---
updated-dependencies:
- dependency-name: mypy
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <[email protected]>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2024-04-25 09:54:50 +05:45
skshetry 8517cccff4 artifacts: add support for monorepo (#10386) 2024-04-16 21:13:02 +05:45
dberenbaum c52a25e011 fix tests 2024-04-16 11:15:52 -04:00
dberenbaum c4678736b0 add option to not relink 2024-04-16 11:15:52 -04:00
pre-commit-ci[bot] ec92cc14d1 [pre-commit.ci] pre-commit autoupdate (#10390)
updates:
- [github.com/astral-sh/ruff-pre-commit: v0.3.5 → v0.3.7](https://github.com/astral-sh/ruff-pre-commit/compare/v0.3.5...v0.3.7)

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2024-04-16 10:37:51 +05:45
Dave Berenbaum e6048287cb fix test for dvc-render updates to plots title anchor (#10387) 2024-04-11 19:23:03 +05:45
pre-commit-ci[bot] 72a353d8b9 [pre-commit.ci] pre-commit autoupdate (#10384)
updates:
- [github.com/pre-commit/pre-commit-hooks: v4.5.0 → v4.6.0](https://github.com/pre-commit/pre-commit-hooks/compare/v4.5.0...v4.6.0)

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2024-04-09 06:45:19 +05:45
pre-commit-ci[bot] dd2b515c97 [pre-commit.ci] pre-commit autoupdate (#10376)
updates:
- [github.com/astral-sh/ruff-pre-commit: v0.3.4 → v0.3.5](https://github.com/astral-sh/ruff-pre-commit/compare/v0.3.4...v0.3.5)

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2024-04-02 06:52:09 +05:45
skshetry c223cdaecd params/metrics: do not fail during dir expansion (#10375)
params/metrics: assume file when dir expansion raises error

Closes https://github.com/iterative/dvc/issues/10018.

We have to be careful when expanding paths. This PR fall backs to assume it
as a file when either of `isdir` and `find` raises error due to some mishaps.
The consequence of this assumption happens when we try to read the file.
In that case, if the directory exists, and we happen to try to open it
as a file, dvc will raise `IsADirectoryError`, but that is handled correctly
and won't fail catastrophically like this did.
2024-03-28 07:30:12 +05:45
skshetry 5379aa2348 add: raise DoesNotExistError only if the output path does not exist (#10373)
`dvc add` command incorrectly raises a `DoesNotExistError` when a
broken symlink exists in an output directory, and the target name
is same as the directory's name.

eg: If `data` is an output, and is the command is invoked as `dvc add data`
(i.e. no virtual directory operations to perform).

The expected behavior to raise a `FileNotFoundError`.
`DoesNotExistError` should only be raised if the output itself does not
exist.

Related: https://github.com/iterative/dvc/issues/3717
2024-03-26 11:16:35 +00:00
pre-commit-ci[bot]andSaugat Pachhai 3375a7c39f [pre-commit.ci] pre-commit autoupdate (#10372)
* [pre-commit.ci] pre-commit autoupdate

updates:
- [github.com/astral-sh/ruff-pre-commit: v0.3.3 → v0.3.4](https://github.com/astral-sh/ruff-pre-commit/compare/v0.3.3...v0.3.4)

* fixes

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Saugat Pachhai (सौगात) <[email protected]>
2024-03-26 06:37:53 +00:00
dberenbaum 519c47f5fc sync: skip if remote doesn't match output remote 2024-03-25 15:12:33 -04:00
skshetry 6f6038888f sort imports for all libraries starting with dvc_ as a first party (#10371)
dvc_studio_client, dvc_ssh, etc. were not properly sorted.
2024-03-23 20:13:13 +05:45
skshetry e223f51e7e explicitly specify dependencies that are imported inside dvc (#10370)
Most of these are also transitive dependencies, but we should explicitly
specify them if they are imported within dvc.
2024-03-23 18:51:25 +05:45
skshetry fb628bf9ea add a regression test for #10309 (#10369)
https://github.com/iterative/dvc/issues/10309
2024-03-23 11:37:36 +05:45
dependabot[bot]andskshetry 4f4e977e53 build(deps-dev): update dvc-data requirement from <3.15,>=3.13 to >=3.15,<3.16 (#10368)
* build(deps-dev): update dvc-data requirement

Updates the requirements on [dvc-data](https://github.com/iterative/dvc-data) to permit the latest version.
- [Release notes](https://github.com/iterative/dvc-data/releases)
- [Commits](https://github.com/iterative/dvc-data/compare/3.13.0...3.15.0)

---
updated-dependencies:
- dependency-name: dvc-data
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <[email protected]>

* Update pyproject.toml

---------

Signed-off-by: dependabot[bot] <[email protected]>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: skshetry <[email protected]>
2024-03-23 05:48:17 +00:00
skshetry 8b1f8847fe add CITATION file so that users can cite dvc (#10366) 2024-03-23 07:05:07 +05:45
skshetry c62a0380e9 templating: support on top-level params/metrics/plots/artifacts/datasets (#10359)
* templating: support on top-level params/metrics/plots/artifacts/datasets

* make resolving top-level ordered

* prevent updating parametrized dataset

* fix dataset tests

* add tests

* fix comment

* support parametrizing artifacts name
2024-03-22 12:22:16 +00:00
Rishabh SharmaandRishabh Sharma 8d8939fb67 DVC_STAGE: Added environment variable to track DVC stage name (#10357)
DVC_STAGE_NAME: Added environment variable to track DVC stage name

Fixes #10355

Co-authored-by: Rishabh Sharma <[email protected]>
2024-03-20 13:28:02 -07:00
dependabot[bot] f7a54bf2d0 build(deps): bump rtCamp/action-slack-notify from 2.2.1 to 2.3.0 (#10364)
Bumps [rtCamp/action-slack-notify](https://github.com/rtcamp/action-slack-notify) from 2.2.1 to 2.3.0.
- [Release notes](https://github.com/rtcamp/action-slack-notify/releases)
- [Commits](https://github.com/rtcamp/action-slack-notify/compare/v2.2.1...v2.3.0)

---
updated-dependencies:
- dependency-name: rtCamp/action-slack-notify
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <[email protected]>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2024-03-20 08:15:27 +05:45
pre-commit-ci[bot]andSaugat Pachhai 1a9f5d3aa8 [pre-commit.ci] pre-commit autoupdate (#10362)
* [pre-commit.ci] pre-commit autoupdate

updates:
- [github.com/astral-sh/ruff-pre-commit: v0.3.2 → v0.3.3](https://github.com/astral-sh/ruff-pre-commit/compare/v0.3.2...v0.3.3)

* fix lint issue

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Saugat Pachhai (सौगात) <[email protected]>
2024-03-19 13:25:39 +05:45
Ivan Shcheklein 017a510b83 add attrs explicit dependency (#10354)
fixes AttrsInstance import failing, fixed by https://github.com/python-attrs/attrs/issues/987
2024-03-16 08:58:37 +05:45
skshetry a80d224c58 fix dvcx import (#10351) 2024-03-14 15:20:17 +05:45
pre-commit-ci[bot]andSaugat Pachhai 49709a4810 [pre-commit.ci] pre-commit autoupdate (#10350)
* [pre-commit.ci] pre-commit autoupdate

updates:
- [github.com/astral-sh/ruff-pre-commit: v0.2.2 → v0.3.2](https://github.com/astral-sh/ruff-pre-commit/compare/v0.2.2...v0.3.2)

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* fix ruff errors

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Saugat Pachhai (सौगात) <[email protected]>
2024-03-13 16:43:10 +05:45
dberenbaum 7df58ce404 Revert "refactor notify_refs_to_studio"
This reverts commit c16d615b55f2bb359b731939ad22712a502d8a8d.
2024-03-11 10:47:01 -04:00
dberenbaum 827e58f305 Revert "fix mypy"
This reverts commit 029b68ef77b7f82d5e72c8bb48f3eb980fe8204e.
2024-03-11 10:47:01 -04:00
dberenbaum 8bc6ba7cd5 fix mypy 2024-03-11 10:47:01 -04:00
dberenbaum a96a446742 refactor notify_refs_to_studio 2024-03-11 10:47:01 -04:00
dberenbaum 837c8e8de4 exp: use exp git remote for studio repo url 2024-03-11 10:47:01 -04:00
dberenbaum 603da1ca1e api: update remote config tests 2024-03-11 09:47:05 -04:00
dberenbaum 377d0c9d58 get: update show-url to take remote args 2024-03-11 09:47:05 -04:00
dependabot[bot] 6bdf41a217 build(deps-dev): bump mypy from 1.8.0 to 1.9.0 (#10348)
Bumps [mypy](https://github.com/python/mypy) from 1.8.0 to 1.9.0.
- [Changelog](https://github.com/python/mypy/blob/master/CHANGELOG.md)
- [Commits](https://github.com/python/mypy/compare/v1.8.0...1.9.0)

---
updated-dependencies:
- dependency-name: mypy
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <[email protected]>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
2024-03-11 08:44:35 +05:45
dberenbaum 1aeb8750b6 drop debugging comment 2024-03-07 07:39:53 -05:00
dberenbaum 5aae934379 test auto push with tmp dir 2024-03-07 07:39:53 -05:00
dberenbaum de1338caaf write git config to exp tmp repos 2024-03-07 07:39:53 -05:00
Ivan Shchekleinandskshetry 9957410998 fix(plots): handle empty directories (#10344)
* fix(plots): handle empty directories

* Update tests/integration/plots/test_plots.py

Co-authored-by: skshetry <[email protected]>

* cleanup after review

---------

Co-authored-by: skshetry <[email protected]>
2024-03-07 12:50:49 +05:45
dberenbaum cf9855af45 handle auto push errors 2024-03-05 09:44:48 -05:00
skshetry 0001cc3a4b drop 2.x branch (#10339)
* drop 2.x branch

* Update tests.yaml
2024-03-05 14:39:51 +05:45
skshetry 440e1491f0 Delete .github/mergify.yml (#10338) 2024-03-05 14:37:44 +05:45
AlexandreKempfandSaugat Pachhai b0384f8f51 dvc studio login automatically push experiments by default (#10331)
* studio login set auto_push + add logs for the user

* set a separate global config dir for studio login tests

The `isolate` fixture sets a separate, _isolated_ global config directory.
But it is shared among other tests. So any mutations to it affects other
tests too.

We should force `studio login` tests to use a separate tmp_dir for global config.

* cleanup message

---------

Co-authored-by: Saugat Pachhai (सौगात) <[email protected]>
2024-03-05 11:48:16 +05:45
108 changed files with 1586 additions and 301 deletions
-10
View File
@@ -1,10 +0,0 @@
pull_request_rules:
- name: backport patches to 2.x
conditions:
- base=main
- label=backport
actions:
backport:
branches:
- "2.x"
title: ({{ destination_branch }}) {{ title }}
+1 -1
View File
@@ -2,7 +2,7 @@ name: "CodeQL"
on:
push:
branches: ["main", "2.x"]
branches: ["main"]
pull_request:
branches: ["main"]
schedule:
+2 -2
View File
@@ -2,7 +2,7 @@ name: Tests
on:
push:
branches: [main, 2.x]
branches: [main]
pull_request:
schedule:
- cron: '5 1 * * *' # every day at 01:05
@@ -66,7 +66,7 @@ jobs:
runs-on: ubuntu-latest
steps:
- name: Slack Notification
uses: rtCamp/action-slack-notify@v2.2.1
uses: rtCamp/action-slack-notify@v2.3.0
env:
SLACK_WEBHOOK: ${{ secrets.SLACK_WEBHOOK }}
SLACK_COLOR: ${{ job.status }}
+2 -2
View File
@@ -6,7 +6,7 @@ ci:
repos:
- repo: https://github.com/pre-commit/pre-commit-hooks
rev: v4.5.0
rev: v4.6.0
hooks:
- id: check-added-large-files
- id: check-case-conflict
@@ -22,7 +22,7 @@ repos:
- id: sort-simple-yaml
- id: trailing-whitespace
- repo: https://github.com/astral-sh/ruff-pre-commit
rev: 'v0.2.2'
rev: 'v0.4.4'
hooks:
- id: ruff
args: [--fix, --exit-non-zero-on-fix]
+25
View File
@@ -0,0 +1,25 @@
cff-version: 1.2.0
title: 'DVC: Data Version Control - Git for Data & Models"'
message: >-
If you use this software as part of a publication and wish
to cite it, please use the metadata from this file.
type: software
authors:
- name: The DVC team and contributors
website: 'https://github.com/iterative'
repository-code: 'https://github.com/iterative/dvc'
url: 'https://dvc.org'
abstract: >-
DVC is a tool for data science that takes advantage of existing software engineering toolset.
It helps machine learning teams manage large datasets, make projects reproducible, and collaborate better.
keywords:
- ai
- collaboration
- data-science
- data-version-control
- developer-tools
- git
- machine-learning
- python
- reproducibility
license: Apache-2.0
+1
View File
@@ -3,6 +3,7 @@ DVC
----
Make your data science projects reproducible and shareable.
"""
import dvc.logger
from dvc.build import PKG # noqa: F401
from dvc.version import __version__, version_tuple # noqa: F401
+1
View File
@@ -1,4 +1,5 @@
"""Main entry point for DVC command line tool."""
import sys
from dvc.cli import main
+1 -1
View File
@@ -162,7 +162,7 @@ def _get_path_func(tool: str, ext: str):
fmt = f"{tool}.dvc-{{now:%Y%m%d}}_{{now:%H%M%S}}.{ext}"
def func(now: Optional["datetime"] = None) -> str:
return fmt.format(now=now or datetime.now())
return fmt.format(now=now or datetime.now()) # noqa: DTZ005
return func
+13 -2
View File
@@ -1,3 +1,4 @@
import os
from typing import Any, Optional
from dvc.repo import Repo
@@ -36,12 +37,22 @@ def artifacts_show(
if version and stage:
raise ValueError("Artifact version and stage are mutually exclusive.")
from dvc.repo.artifacts import Artifacts
from dvc.utils import as_posix
repo_kwargs: dict[str, Any] = {
"subrepos": True,
"uninitialized": True,
}
dirname, _ = Artifacts.parse_path(name)
with Repo.open(repo, **repo_kwargs) as _repo:
rev = _repo.artifacts.get_rev(name, version=version, stage=stage)
with _repo.switch(rev):
path = _repo.artifacts.get_path(name)
return {"rev": rev, "path": path}
root = _repo.fs.root_marker
_dirname = _repo.fs.join(root, dirname) if dirname else root
with Repo(_dirname, fs=_repo.fs, scm=_repo.scm) as r:
path = r.artifacts.get_path(name)
path = _repo.fs.join(_repo.fs.root_marker, as_posix(path))
parts = _repo.fs.relparts(path, _repo.root_dir)
return {"rev": rev, "path": os.path.join(*parts)}
+41 -5
View File
@@ -25,7 +25,14 @@ def _wrap_exceptions(repo, url):
raise PathMissingError(exc.path, url) from exc
def get_url(path, repo=None, rev=None, remote=None):
def get_url(
path: str,
repo: Optional[str] = None,
rev: Optional[str] = None,
remote: Optional[str] = None,
config: Optional[dict[str, Any]] = None,
remote_config: Optional[dict[str, Any]] = 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
@@ -35,15 +42,44 @@ def get_url(path, repo=None, rev=None, remote=None):
NOTE: This function does not check for the actual existence of the file or
directory in the remote storage.
Args:
path (str): location and file name of the target, 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.
config(dict, optional): config to be passed to the DVC repository.
Defaults to None.
remote_config(dict, optional): remote config to be passed to the DVC
repository.
Defaults to None.
Returns:
str: URL to the file or directory.
"""
from dvc.config import NoRemoteError
from dvc_data.index import StorageKeyError
repo_kwargs: dict[str, Any] = {}
if remote:
repo_kwargs["config"] = {"core": {"remote": remote}}
with Repo.open(
repo, rev=rev, subrepos=True, uninitialized=True, **repo_kwargs
repo,
rev=rev,
subrepos=True,
uninitialized=True,
remote=remote,
config=config,
remote_config=remote_config,
) as _repo:
index, entry = _repo.get_data_index_entry(path)
with reraise(
+1
View File
@@ -1,4 +1,5 @@
"""Main parser for the dvc cli."""
import argparse
import os
from functools import lru_cache
+8
View File
@@ -51,6 +51,7 @@ class CmdAdd(CmdBase):
to_remote=self.args.to_remote,
remote_jobs=self.args.remote_jobs,
force=self.args.force,
relink=self.args.relink,
)
except FileNotFoundError:
logger.exception("")
@@ -119,6 +120,13 @@ def add_parser(subparsers, parent_parser):
default=False,
help="Override local file or folder if exists.",
)
parser.add_argument(
"--no-relink",
dest="relink",
action="store_false",
help="Don't recreate links from cache to workspace.",
)
parser.set_defaults(relink=True)
parser.add_argument(
"targets", nargs="+", help="Input files/directories to add."
).complete = completion.FILE
+33
View File
@@ -13,6 +13,9 @@ class CmdArtifactsGet(CmdBaseNoRepo):
from dvc.scm import CloneError
from dvc.ui import ui
if self.args.show_url:
return self._show_url()
try:
count, out = Artifacts.get(
self.args.url,
@@ -36,6 +39,28 @@ class CmdArtifactsGet(CmdBaseNoRepo):
)
return 1
def _show_url(self):
from dvc.api import artifacts_show, get_url
from dvc.ui import ui
artifact = artifacts_show(
self.args.name,
version=self.args.rev,
stage=self.args.stage,
repo=self.args.url,
)
url = get_url(
artifact["path"],
repo=self.args.url,
rev=artifact["rev"],
remote=self.args.remote,
remote_config=self.args.remote_config,
)
ui.write(url, force=True)
return 0
def add_parser(subparsers, parent_parser):
ARTIFACTS_HELP = "DVC model registry artifact commands."
@@ -84,6 +109,14 @@ def add_parser(subparsers, parent_parser):
help="Destination path to download artifact to",
metavar="<path>",
).complete = completion.DIR
get_parser.add_argument(
"--show-url",
action="store_true",
help=(
"Print the storage location (URL) the target data would be "
"downloaded from, and exit."
),
)
get_parser.add_argument(
"-j",
"--jobs",
+8
View File
@@ -20,6 +20,7 @@ class CmdCommit(CmdBase):
with_deps=self.args.with_deps,
recursive=self.args.recursive,
force=self.args.force,
relink=self.args.relink,
)
except DvcException:
logger.exception("failed to commit%s", (" " + target) if target else "")
@@ -64,6 +65,13 @@ def add_parser(subparsers, parent_parser):
default=False,
help="Commit cache for subdirectories of the specified directory.",
)
commit_parser.add_argument(
"--no-relink",
dest="relink",
action="store_false",
help="Don't recreate links from cache to workspace.",
)
commit_parser.set_defaults(relink=True)
commit_parser.add_argument(
"targets",
nargs="*",
+6 -6
View File
@@ -194,8 +194,8 @@ def add_parser(subparsers, _parent_parser):
)
pull_parser.add_argument(
"--run-cache",
action="store_true",
default=False,
action=argparse.BooleanOptionalAction,
default=True,
help="Fetch run history for all stages.",
)
pull_parser.add_argument(
@@ -262,8 +262,8 @@ def add_parser(subparsers, _parent_parser):
)
push_parser.add_argument(
"--run-cache",
action="store_true",
default=False,
action=argparse.BooleanOptionalAction,
default=True,
help="Push run history for all stages.",
)
push_parser.add_argument(
@@ -324,8 +324,8 @@ def add_parser(subparsers, _parent_parser):
)
fetch_parser.add_argument(
"--run-cache",
action="store_true",
default=False,
action=argparse.BooleanOptionalAction,
default=True,
help="Fetch run history for all stages.",
)
fetch_parser.add_argument(
+4 -2
View File
@@ -1,3 +1,5 @@
import argparse
from dvc.cli import formatter
from dvc.cli.command import CmdBase
from dvc.cli.utils import append_doc_link
@@ -81,8 +83,8 @@ def add_parser(experiments_subparsers, parent_parser):
)
experiments_pull_parser.add_argument(
"--run-cache",
action="store_true",
default=False,
action=argparse.BooleanOptionalAction,
default=True,
help="Pull run history for all stages.",
)
experiments_pull_parser.add_argument(
+3 -2
View File
@@ -1,3 +1,4 @@
import argparse
from typing import Any
from dvc.cli import completion, formatter
@@ -120,8 +121,8 @@ def add_parser(experiments_subparsers, parent_parser):
)
experiments_push_parser.add_argument(
"--run-cache",
action="store_true",
default=False,
action=argparse.BooleanOptionalAction,
default=True,
help="Push run history for all stages.",
)
experiments_push_parser.add_argument(
+1 -1
View File
@@ -60,7 +60,7 @@ class CmdGC(CmdBase):
msg += " of the current and the following repos:"
for repo_path in self.args.repos:
msg += "\n - %s" % os.path.abspath(repo_path)
msg += f"\n - {os.path.abspath(repo_path)}"
else:
msg += " of the current repo."
+7 -1
View File
@@ -12,7 +12,13 @@ class CmdGet(CmdBaseNoRepo):
from dvc.api import get_url
from dvc.ui import ui
url = get_url(self.args.path, repo=self.args.url, rev=self.args.rev)
url = get_url(
self.args.path,
repo=self.args.url,
rev=self.args.rev,
remote=self.args.remote,
remote_config=self.args.remote_config,
)
ui.write(url, force=True)
return 0
+1 -4
View File
@@ -117,10 +117,7 @@ and then the stage name name.
"--pull",
action="store_true",
default=False,
help=(
"Try automatically pulling missing cache for outputs restored "
"from the run-cache."
),
help="Try automatically pulling missing data.",
)
repro_parser.add_argument(
"--allow-missing",
+12 -2
View File
@@ -13,11 +13,10 @@ logger = logger.getChild(__name__)
class CmdStudioLogin(CmdConfig):
def run(self):
from dvc_studio_client.auth import StudioAuthError, get_access_token
from dvc.env import DVC_STUDIO_URL
from dvc.ui import ui
from dvc.utils.studio import STUDIO_URL
from dvc_studio_client.auth import StudioAuthError, get_access_token
name = self.args.name
hostname = self.args.hostname or os.environ.get(DVC_STUDIO_URL) or STUDIO_URL
@@ -46,6 +45,15 @@ class CmdStudioLogin(CmdConfig):
self.save_config(hostname, access_token)
if not self.config["exp"].get("auto_push", True):
from dvc.ui import ui
ui.warn(
"exp.auto_push is disabled. \n"
"Enable with 'dvc config exp.auto_push true' "
"to automatically push experiments to Studio."
)
config_path = self.config.files["global"]
ui.write(f"Authentication complete. Saved token to {config_path}.")
return 0
@@ -54,6 +62,8 @@ class CmdStudioLogin(CmdConfig):
with self.config.edit("global") as conf:
conf["studio"]["token"] = token
conf["studio"]["url"] = hostname
if "auto_push" not in conf["exp"]:
conf["exp"]["auto_push"] = True
class CmdStudioLogout(CmdConfig):
+2 -4
View File
@@ -88,12 +88,10 @@ class TabularData(MutableSequence[Sequence["CellT"]]):
return list(it)
@overload
def __setitem__(self, item: int, value: Sequence["CellT"]) -> None:
...
def __setitem__(self, item: int, value: Sequence["CellT"]) -> None: ...
@overload
def __setitem__(self, item: slice, value: Iterable[Sequence["CellT"]]) -> None:
...
def __setitem__(self, item: slice, value: Iterable[Sequence["CellT"]]) -> None: ...
def __setitem__(self, item, value) -> None:
it = value
+1
View File
@@ -1,4 +1,5 @@
"""DVC config objects."""
import os
import re
from contextlib import contextmanager
+7 -1
View File
@@ -193,6 +193,7 @@ SCHEMA = {
"session_token": str,
Optional("listobjects", default=False): Bool, # obsoleted
Optional("use_ssl", default=True): Bool,
Optional("allow_anonymous_login", default=False): Bool,
"ssl_verify": Any(Bool, str),
"sse": str,
"sse_kms_key_id": str,
@@ -232,7 +233,12 @@ SCHEMA = {
Optional("verify", default=False): Bool,
**REMOTE_COMMON,
},
"hdfs": {"user": str, "kerb_ticket": str, **REMOTE_COMMON},
"hdfs": {
"user": str,
"kerb_ticket": str,
"replication": int,
**REMOTE_COMMON,
},
"webhdfs": {
"kerberos": Bool,
"kerberos_principal": str,
+6 -3
View File
@@ -22,9 +22,12 @@ def download_progress(to: "Output") -> Iterator[Callable[[int], Any]]:
from dvc.ui import ui
from dvc.ui._rich_progress import DbDownloadProgress
with log_durations(logger.debug, f"Saving to {to}"), DbDownloadProgress(
console=ui.error_console,
) as progress:
with (
log_durations(logger.debug, f"Saving to {to}"),
DbDownloadProgress(
console=ui.error_console,
) as progress,
):
task = progress.add_task("Saving", total=None, output=to)
yield lambda n: progress.advance(task, advance=n)
progress.update(task, description="Saved", total=0)
+1 -1
View File
@@ -35,7 +35,7 @@ class RepoDependency(Dependency):
super().__init__(stage, *args, **kwargs)
self.fs = self._make_fs()
self.fs_path = as_posix(self.def_path)
self.fs_path = as_posix(self.fs.normpath(self.def_path))
def _parse_path(self, fs, fs_path): # noqa: ARG002
return None
+24 -19
View File
@@ -68,10 +68,9 @@ def is_git_ignored(repo, path):
def check_dvcfile_path(repo, path):
if not is_valid_filename(path):
raise StageFileBadNameError(
"bad DVC file name '{}'. DVC files should be named "
"'{}' or have a '.dvc' suffix (e.g. '{}.dvc').".format(
relpath(path), PROJECT_FILE, os.path.basename(path)
)
f"bad DVC file name '{relpath(path)}'. DVC files should be named "
f"'{PROJECT_FILE}' or have a '.dvc' suffix "
f"(e.g. '{os.path.basename(path)}.dvc')."
)
if is_git_ignored(repo, path):
@@ -243,16 +242,22 @@ class ProjectFile(FileMixin):
def dump_dataset(self, dataset):
with modify_yaml(self.path, fs=self.repo.fs) as data:
datasets: list[dict] = data.setdefault("datasets", [])
parsed = self.datasets if data else []
raw = data.setdefault("datasets", [])
loc = next(
(i for i, ds in enumerate(datasets) if ds["name"] == dataset["name"]),
(i for i, ds in enumerate(parsed) if ds["name"] == dataset["name"]),
None,
)
if loc is not None:
apply_diff(dataset, datasets[loc])
datasets[loc] = dataset
if raw[loc] != parsed[loc]:
raise ParametrizedDumpError(
"cannot update a parametrized dataset entry"
)
apply_diff(dataset, raw[loc])
raw[loc] = dataset
else:
datasets.append(dataset)
raw.append(dataset)
self.repo.scm_context.track_file(self.relpath)
def _dump_lockfile(self, stage, **kwargs):
@@ -308,29 +313,29 @@ class ProjectFile(FileMixin):
return self.LOADER(self, self.contents, self.lockfile_contents)
@property
def metrics(self) -> list[str]:
return self.contents.get("metrics", [])
def artifacts(self) -> dict[str, Optional[dict[str, Any]]]:
return self.resolver.resolve_artifacts()
@property
def plots(self) -> Any:
return self.contents.get("plots", {})
def metrics(self) -> list[str]:
return self.resolver.resolve_metrics()
@property
def params(self) -> list[str]:
return self.contents.get("params", [])
return self.resolver.resolve_params()
@property
def plots(self) -> list[Any]:
return self.resolver.resolve_plots()
@property
def datasets(self) -> list[dict[str, Any]]:
return self.contents.get("datasets", [])
return self.resolver.resolve_datasets()
@property
def datasets_lock(self) -> list[dict[str, Any]]:
return self.lockfile_contents.get("datasets", [])
@property
def artifacts(self) -> dict[str, Optional[dict[str, Any]]]:
return self.contents.get("artifacts", {})
def remove(self, force=False):
if not force:
logger.warning("Cannot remove pipeline file.")
+1
View File
@@ -19,3 +19,4 @@ DVC_STUDIO_URL = "DVC_STUDIO_URL"
DVC_SQLALCHEMY_ECHO = "DVC_SQLALCHEMY_ECHO"
DVC_SYSTEM_CONFIG_DIR = "DVC_SYSTEM_CONFIG_DIR"
DVC_UPDATER_ENDPOINT = "DVC_UPDATER_ENDPOINT"
DVC_STAGE = "DVC_STAGE"
+6 -6
View File
@@ -1,4 +1,5 @@
"""Exceptions raised by the dvc."""
import errno
from typing import TYPE_CHECKING, Optional
@@ -132,11 +133,11 @@ class MoveNotDataSourceError(DvcException):
def __init__(self, path):
msg = (
"move is not permitted for stages that are not data sources. "
"You need to either move '{path}' to a new location and edit "
"it by hand, or remove '{path}' and create a new one at the "
f"You need to either move {path!r} to a new location and edit "
f"it by hand, or remove {path!r} and create a new one at the "
"desired location."
)
super().__init__(msg.format(path=path))
super().__init__(msg)
class NotDvcRepoError(DvcException):
@@ -270,9 +271,8 @@ class NoOutputInExternalRepoError(DvcException):
from dvc.utils import relpath
super().__init__(
"Output '{}' not found in target repository '{}'".format(
relpath(path, external_repo_path), external_repo_url
)
f"Output {relpath(path, external_repo_path)!r} "
f"not found in target repository '{external_repo_url}'"
)
+10 -3
View File
@@ -1,10 +1,10 @@
import glob
from typing import Optional
from urllib.parse import urlparse
from dvc_http import HTTPFileSystem, HTTPSFileSystem # noqa: F401
from dvc.config import ConfigError as RepoConfigError
from dvc.config_schema import SCHEMA, Invalid
from dvc_http import HTTPFileSystem, HTTPSFileSystem # noqa: F401
# pylint: disable=unused-import
from dvc_objects.fs import ( # noqa: F401
@@ -70,7 +70,14 @@ def download(
to_infos = [to]
if isinstance(fs, DVCFileSystem):
lfs_prefetch(fs, from_infos)
lfs_prefetch(
fs,
[
f"{fs.normpath(glob.escape(fs_path))}/**"
if fs.isdir(fs_path)
else glob.escape(fs_path)
],
)
cb.set_size(len(from_infos))
jobs = jobs or fs.jobs
generic.copy(fs, from_infos, localfs, to_infos, callback=cb, batch_size=jobs)
+1 -1
View File
@@ -360,7 +360,7 @@ class _DVCFileSystem(AbstractFileSystem):
except FileNotFoundError:
return False
def ls(self, path, detail=True, dvc_only=False, **kwargs): # noqa: C901
def ls(self, path, detail=True, dvc_only=False, **kwargs): # noqa: C901, PLR0912
key = self._get_key_from_relative(path)
repo, dvc_fs, subkey = self._get_subrepo_info(key)
+1 -1
View File
@@ -1365,7 +1365,7 @@ class Output:
dry_run=not self.use_cache,
)
except FileNotFoundError as exc:
if self.fs_path == path:
if not self.exists:
raise self.DoesNotExistError(self) from exc
if not self.is_dir_checksum:
raise
+90 -1
View File
@@ -21,6 +21,7 @@ from .context import (
VarsAlreadyLoaded,
)
from .interpolate import (
check_expression,
check_recursive_parse_errors,
is_interpolated_string,
recurse,
@@ -38,10 +39,16 @@ if TYPE_CHECKING:
logger = logger.getChild(__name__)
STAGES_KWD = "stages"
VARS_KWD = "vars"
WDIR_KWD = "wdir"
ARTIFACTS_KWD = "artifacts"
DATASETS_KWD = "datasets"
METRICS_KWD = "metrics"
PARAMS_KWD = "params"
PLOTS_KWD = "plots"
STAGES_KWD = "stages"
FOREACH_KWD = "foreach"
MATRIX_KWD = "matrix"
DO_KWD = "do"
@@ -163,6 +170,27 @@ class DataResolver:
for name, definition in stages_data.items()
}
self.artifacts = [
ArtifactDefinition(self, self.context, name, definition, ARTIFACTS_KWD)
for name, definition in d.get(ARTIFACTS_KWD, {}).items()
]
self.datasets = [
TopDefinition(self, self.context, str(i), definition, DATASETS_KWD)
for i, definition in enumerate(d.get(DATASETS_KWD, []))
]
self.metrics = [
TopDefinition(self, self.context, str(i), definition, METRICS_KWD)
for i, definition in enumerate(d.get(METRICS_KWD, []))
]
self.params = [
TopDefinition(self, self.context, str(i), definition, PARAMS_KWD)
for i, definition in enumerate(d.get(PARAMS_KWD, []))
]
self.plots = [
TopDefinition(self, self.context, str(i), definition, PLOTS_KWD)
for i, definition in enumerate(d.get(PLOTS_KWD, []))
]
def resolve_one(self, name: str):
group, key = split_group_name(name)
@@ -186,6 +214,27 @@ class DataResolver:
logger.trace("Resolved dvc.yaml:\n%s", data)
return {STAGES_KWD: data}
# Top-level sections are eagerly evaluated, whereas stages are lazily evaluated,
# one-by-one.
def resolve_artifacts(self) -> dict[str, Optional[dict[str, Any]]]:
d: dict[str, Optional[dict[str, Any]]] = {}
for item in self.artifacts:
d.update(item.resolve())
return d
def resolve_datasets(self) -> list[dict[str, Any]]:
return [item.resolve() for item in self.datasets]
def resolve_metrics(self) -> list[str]:
return [item.resolve() for item in self.metrics]
def resolve_params(self) -> list[str]:
return [item.resolve() for item in self.params]
def resolve_plots(self) -> list[Any]:
return [item.resolve() for item in self.plots]
def has_key(self, key: str):
return self._has_group_and_key(*split_group_name(key))
@@ -565,3 +614,43 @@ class MatrixDefinition:
return entry.resolve_stage(skip_checks=True)
except ContextError as exc:
format_and_raise(exc, f"stage '{generated}'", self.relpath)
class TopDefinition:
def __init__(
self,
resolver: DataResolver,
context: Context,
name: str,
definition: "Any",
where: str,
):
self.resolver = resolver
self.context = context
self.name = name
self.definition = definition
self.where = where
self.relpath = self.resolver.relpath
def resolve(self):
try:
check_recursive_parse_errors(self.definition)
return self.context.resolve(self.definition)
except (ParseError, ContextError) as exc:
format_and_raise(exc, f"'{self.where}.{self.name}'", self.relpath)
class ArtifactDefinition(TopDefinition):
def resolve(self) -> dict[str, Optional[dict[str, Any]]]:
try:
check_expression(self.name)
name = self.context.resolve(self.name)
if not isinstance(name, str):
typ = type(name).__name__
raise ResolveError(
f"failed to resolve '{self.where}.{self.name}'"
f" in '{self.relpath}': expected str, got " + typ
)
except (ParseError, ContextError) as exc:
format_and_raise(exc, f"'{self.where}.{self.name}'", self.relpath)
return {name: super().resolve()}
+1
View File
@@ -1,4 +1,5 @@
"""Manages progress bars for DVC repo."""
import logging
import sys
from threading import RLock
+1 -5
View File
@@ -551,11 +551,7 @@ class Repo:
if out.protocol == "local" and match(fs_path, out.fs_path):
return True
if recursive and out.fs.isin(out.fs_path, fs_path):
return True
return False
return recursive and out.fs.isin(out.fs_path, fs_path)
matched = list(filter(func, outs))
if not matched:
+14 -3
View File
@@ -171,11 +171,16 @@ def _add_transfer(
stage.dump()
def _add(stage: "Stage", source: Optional[str] = None, no_commit: bool = False) -> None:
def _add(
stage: "Stage",
source: Optional[str] = None,
no_commit: bool = False,
relink: bool = True,
) -> None:
out = stage.outs[0]
path = out.fs.abspath(source) if source else None
try:
stage.add_outs(path, no_commit=no_commit)
stage.add_outs(path, no_commit=no_commit, relink=relink)
except CacheLinkError:
stage.dump()
raise
@@ -194,6 +199,7 @@ def add(
to_remote: bool = False,
remote_jobs: Optional[int] = None,
force: bool = False,
relink: bool = True,
) -> list["Stage"]:
add_targets = find_targets(targets, glob=glob)
if not add_targets:
@@ -224,7 +230,12 @@ def add(
with warn_link_failures() as link_failures:
for source, (stage, output_exists) in progress_iter(stages_with_targets):
try:
_add(stage, source if output_exists else None, no_commit=no_commit)
_add(
stage,
source if output_exists else None,
no_commit=no_commit,
relink=relink,
)
except CacheLinkError:
link_failures.append(stage.relpath)
return stages
+63 -19
View File
@@ -12,7 +12,7 @@ from dvc.exceptions import (
InvalidArgumentError,
)
from dvc.log import logger
from dvc.utils import relpath, resolve_output
from dvc.utils import as_posix, relpath, resolve_output
from dvc.utils.objects import cached_property
from dvc.utils.serialize import modify_yaml
@@ -99,7 +99,7 @@ class Artifacts:
"""Read artifacts from dvc.yaml."""
artifacts: dict[str, dict[str, Artifact]] = {}
for dvcfile, dvcfile_artifacts in self.repo.index._artifacts.items():
dvcyaml = relpath(dvcfile, self.repo.root_dir)
dvcyaml = self.repo.fs.relpath(dvcfile, self.repo.root_dir)
artifacts[dvcyaml] = {}
for name, value in dvcfile_artifacts.items():
try:
@@ -147,8 +147,8 @@ class Artifacts:
gto_tags: list["GTOTag"] = sort_versions(parse_tag(tag) for tag in tags)
return gto_tags[0].tag.target
def get_path(self, name: str):
"""Return repo fspath for the given artifact."""
@classmethod
def parse_path(cls, name: str) -> tuple[Optional[str], str]:
from gto.constants import SEPARATOR_IN_NAME, fullname_re
name = _reformat_name(name)
@@ -158,13 +158,37 @@ class Artifacts:
dirname = m.group("dirname")
if dirname:
dirname = dirname.rstrip(SEPARATOR_IN_NAME)
dvcyaml = os.path.join(dirname, PROJECT_FILE) if dirname else PROJECT_FILE
artifact_name = m.group("name")
return dirname, m.group("name")
def get_path(self, name: str):
"""Return fspath for the given artifact relative to the git root."""
from dvc.fs import GitFileSystem
dirname, artifact_name = self.parse_path(name)
# `name`/`dirname` are expected to be a git root relative.
# We convert it to dvc-root relative path so that we can read artifacts
# from dvc.yaml file.
# But we return dirname intact, as we want to return a git-root relative path.
# This is useful when reading from `dvcfs` from remote.
fs = self.repo.fs
assert self.scm
if isinstance(fs, GitFileSystem):
scm_root = fs.root_marker
else:
scm_root = self.scm.root_dir
dirparts = posixpath.normpath(dirname).split(posixpath.sep) if dirname else ()
abspath = fs.join(scm_root, *dirparts, PROJECT_FILE)
rela = fs.relpath(abspath, self.repo.root_dir)
try:
artifact = self.read()[dvcyaml][artifact_name]
artifact = self.read()[rela][artifact_name]
except KeyError as exc:
raise ArtifactNotFoundError(name) from exc
return os.path.join(dirname, artifact.path) if dirname else artifact.path
path = posixpath.join(dirname or "", artifact.path)
parts = posixpath.normpath(path).split(posixpath.sep)
return os.path.join(*parts)
def download(
self,
@@ -177,15 +201,28 @@ class Artifacts:
) -> tuple[int, str]:
"""Download the specified artifact."""
from dvc.fs import download as fs_download
from dvc.repo import Repo
logger.debug("Trying to download artifact '%s' via DVC", name)
rev = self.get_rev(name, version=version, stage=stage)
dirname, _ = self.parse_path(name)
with self.repo.switch(rev):
path = self.get_path(name)
root = self.repo.fs.root_marker
_dirname = self.repo.fs.join(root, dirname) if dirname else root
with Repo(_dirname, fs=self.repo.fs, scm=self.repo.scm) as r:
path = r.artifacts.get_path(name)
path = self.repo.fs.join(root, as_posix(path))
path = self.repo.fs.relpath(path, self.repo.root_dir)
# when the `repo` is a subrepo, the path `/subrepo/myart.pkl` for dvcfs
# should be translated as `/myart.pkl`,
# i.e. relative to the root of the subrepo
path = self.repo.fs.join(root, path)
path = self.repo.fs.normpath(path)
out = resolve_output(path, out, force=force)
fs = self.repo.dvcfs
fs_path = fs.from_os_path(path)
count = fs_download(fs, fs_path, os.path.abspath(out), jobs=jobs)
count = fs_download(fs, path, os.path.abspath(out), jobs=jobs)
return count, out
@staticmethod
@@ -200,10 +237,9 @@ class Artifacts:
dvc_studio_config: Optional[dict[str, Any]] = None,
**kwargs,
) -> tuple[int, str]:
from dvc_studio_client.model_registry import get_download_uris
from dvc.fs import HTTPFileSystem, generic, localfs
from dvc.fs.callbacks import TqdmCallback
from dvc_studio_client.model_registry import get_download_uris
logger.debug("Trying to download artifact '%s' via studio", name)
out = out or os.getcwd()
@@ -231,7 +267,7 @@ class Artifacts:
from_infos.append(url)
except DvcException:
raise
except Exception as exc: # noqa: BLE001
except Exception as exc:
raise DvcException(
f"Failed to download artifact '{name}' via Studio"
) from exc
@@ -278,6 +314,13 @@ class Artifacts:
name = _reformat_name(name)
saved_exc: Optional[Exception] = None
local_dvc_studio_config = Config().get("studio", {})
args_dvc_studio_config = {}
if config and not isinstance(config, dict):
config = Config.load_file(config)
args_dvc_studio_config = config.get("studio", {})
try:
logger.trace("Trying studio-only config")
return cls._download_studio(
@@ -288,14 +331,13 @@ class Artifacts:
out=out,
force=force,
jobs=jobs,
dvc_studio_config=local_dvc_studio_config | args_dvc_studio_config,
)
except FileExistsLocallyError:
raise
except Exception as exc: # noqa: BLE001
saved_exc = exc
if config and not isinstance(config, dict):
config = Config.load_file(config)
with Repo.open(
url=url,
subrepos=True,
@@ -305,7 +347,7 @@ class Artifacts:
remote_config=remote_config,
) as repo:
logger.trace("Trying repo [studio] config")
dvc_studio_config = dict(repo.config.get("studio"))
repo_dvc_studio_config = repo.config.get("studio", {})
try:
return cls._download_studio(
url,
@@ -315,7 +357,9 @@ class Artifacts:
out=out,
force=force,
jobs=jobs,
dvc_studio_config=dvc_studio_config,
dvc_studio_config=local_dvc_studio_config
| repo_dvc_studio_config
| args_dvc_studio_config,
)
except FileExistsLocallyError:
raise
@@ -333,7 +377,7 @@ class Artifacts:
)
except FileExistsLocallyError:
raise
except Exception as exc: # noqa: BLE001
except Exception as exc:
if saved_exc:
logger.exception(str(saved_exc), exc_info=saved_exc.__cause__)
raise DvcException(
+1 -2
View File
@@ -16,8 +16,7 @@ from dvc.types import StrPath
from dvc_data.hashfile.meta import Meta
if TYPE_CHECKING:
from dql.dataset import DatasetRecord # type: ignore[import]
from dvcx.dataset import DatasetVersion # type: ignore[import]
from dvcx.dataset import DatasetRecord, DatasetVersion # type: ignore[import]
from typing_extensions import Self
from dvc.repo import Repo
+2 -2
View File
@@ -78,7 +78,7 @@ def _collect_rev(
timestamp: Optional[datetime] = None
else:
commit = repo.scm.resolve_commit(rev)
timestamp = datetime.fromtimestamp(commit.commit_time)
timestamp = datetime.fromtimestamp(commit.commit_time) # noqa: DTZ006
return SerializableExp.from_repo(
repo,
@@ -338,6 +338,6 @@ def _sorted_ranges(exp_ranges: Iterable["ExpRange"]) -> list["ExpRange"]:
if head_exp and head_exp.data and head_exp.data.timestamp:
return head_exp.data.timestamp, head_exp.rev
return datetime.fromtimestamp(0), ""
return datetime.fromtimestamp(0), "" # noqa: DTZ006
return sorted(exp_ranges, key=_head_timestamp, reverse=True)
+24 -4
View File
@@ -25,6 +25,7 @@ from dvc.utils import dict_sha256, env2bool, relpath
from dvc.utils.fs import remove
from dvc.utils.studio import (
env_to_config,
get_repo_url,
get_subrepo_relpath,
)
@@ -584,9 +585,8 @@ class BaseExecutor(ABC):
message: Optional[str] = None,
**kwargs,
) -> Iterator["Repo"]:
from dvc_studio_client.post_live_metrics import post_live_metrics
from dvc.repo import Repo
from dvc_studio_client.post_live_metrics import post_live_metrics
with Repo(os.path.join(info.root_dir, info.dvc_dir)) as dvc:
info.status = TaskStatus.RUNNING
@@ -611,6 +611,8 @@ class BaseExecutor(ABC):
# inferring repo url will fail if not set here
run_env_config = env_to_config(kwargs.get("run_env", {}))
dvc_studio_config = run_env_config | dvc_studio_config
# override studio repo url if exp git remote set
repo_url = get_repo_url(dvc)
try:
post_live_metrics(
"start",
@@ -621,6 +623,7 @@ class BaseExecutor(ABC):
dvc_studio_config=dvc_studio_config,
message=message,
subdir=get_subrepo_relpath(dvc),
studio_repo_url=repo_url,
)
logger.debug("Running repro in '%s'", os.getcwd())
yield dvc
@@ -646,6 +649,7 @@ class BaseExecutor(ABC):
experiment_rev=dvc.experiments.scm.get_ref(EXEC_BRANCH),
metrics=_gather_metrics(dvc, on_error="return"),
dvc_studio_config=dvc_studio_config,
studio_repo_url=repo_url,
)
if infofile is not None:
@@ -677,7 +681,13 @@ class BaseExecutor(ABC):
git_remote = os.getenv(
DVC_EXP_GIT_REMOTE, exp_config.get("git_remote", "origin")
)
cls._validate_remotes(dvc, git_remote)
try:
cls._validate_remotes(dvc, git_remote)
except DvcException as exc:
logger.warning("Failed to validate remotes. Disabling auto push: %s", exc)
yield
return
yield
cls._auto_push(dvc, git_remote)
@@ -688,7 +698,17 @@ class BaseExecutor(ABC):
push_cache=True,
run_cache=True,
):
from dvc.ui import ui
from dvc.utils import format_link
branch = dvc.scm.get_ref(EXEC_BRANCH, follow=False)
link = format_link(
"https://dvc.org/doc/user-guide/experiment-management/sharing-experiments"
)
ui.write(
f"Pushing experiment to '{git_remote}'. Cancel with CTRL+C. "
f"See {link} for more info."
)
try:
dvc.experiments.push(
git_remote,
@@ -696,7 +716,7 @@ class BaseExecutor(ABC):
push_cache=push_cache,
run_cache=run_cache,
)
except BaseException as exc: # noqa: BLE001
except DvcException as exc:
logger.warning(
(
"Something went wrong while auto pushing experiment "
+18 -3
View File
@@ -96,9 +96,10 @@ class TempDirExecutor(BaseLocalExecutor):
temp_merge: stash_rev,
temp_baseline: entry.baseline_rev,
}
with get_exp_rwlock(
repo, writes=[temp_head, temp_merge, temp_baseline]
), self.set_temp_refs(scm, temp_ref_dict):
with (
get_exp_rwlock(repo, writes=[temp_head, temp_merge, temp_baseline]),
self.set_temp_refs(scm, temp_ref_dict),
):
# Executor will be initialized with an empty git repo that
# we populate by pushing:
# EXEC_HEAD - the base commit for this experiment
@@ -130,6 +131,8 @@ class TempDirExecutor(BaseLocalExecutor):
self.scm.stash.apply(merge_rev)
self._update_config(repo.config.read("local"))
local_git_config = os.path.join(repo.scm.root_dir, ".git", "config")
self._update_git_config(ConfigObj(local_git_config, list_values=False))
def _update_config(self, update):
local_config = os.path.join(self.root_dir, self.dvc_dir, "config.local")
@@ -143,6 +146,18 @@ class TempDirExecutor(BaseLocalExecutor):
with open(local_config, "wb") as fobj:
conf_obj.write(fobj)
def _update_git_config(self, update):
local_config = os.path.join(self.scm.root_dir, ".git", "config")
logger.debug("Writing experiments local Git config '%s'", local_config)
if os.path.exists(local_config):
conf_obj = ConfigObj(local_config, list_values=False)
conf_obj.merge(update)
else:
conf_obj = ConfigObj(update, list_values=False)
if conf_obj:
with open(local_config, "wb") as fobj:
conf_obj.write(fobj)
def init_cache(
self,
repo: "Repo",
+2 -2
View File
@@ -4,7 +4,6 @@ from collections.abc import Collection, Generator, Iterable, Mapping
from dataclasses import asdict, dataclass
from typing import TYPE_CHECKING, Any, NamedTuple, Optional, Union
from dvc_studio_client.post_live_metrics import get_studio_config
from funcy import retry
from dvc.dependency import ParamsDependency
@@ -24,6 +23,7 @@ from dvc.repo.experiments.utils import (
)
from dvc.utils.objects import cached_property
from dvc.utils.studio import config_to_env
from dvc_studio_client.post_live_metrics import get_studio_config
from .utils import get_remote_executor_refs
@@ -180,7 +180,7 @@ class BaseStashQueue(ABC):
def _get_timestamp(rev: str) -> datetime:
commit = self.scm.resolve_commit(rev)
return datetime.fromtimestamp(commit.commit_time)
return datetime.fromtimestamp(commit.commit_time) # noqa: DTZ006
def _format_entry(
entry: QueueEntry,
+4 -4
View File
@@ -474,10 +474,10 @@ class LocalCeleryQueue(BaseStashQueue):
self.scm, exp_names, git_remote
)
if not git_remote:
queue_entry_match: dict[
str, Optional["QueueEntry"]
] = self.match_queue_entry_by_name(
exp_names, self.iter_queued(), self.iter_done()
queue_entry_match: dict[str, Optional["QueueEntry"]] = (
self.match_queue_entry_by_name(
exp_names, self.iter_queued(), self.iter_done()
)
)
for exp_name in exp_names:
+1 -1
View File
@@ -125,7 +125,7 @@ class TempDirQueue(WorkspaceQueue):
)
except DvcException:
raise
except Exception as exc: # noqa: BLE001
except Exception as exc:
raise DvcException(f"Failed to reproduce experiment '{rev[:7]}'") from exc
finally:
executor.cleanup(infofile)
+1 -1
View File
@@ -131,7 +131,7 @@ class WorkspaceQueue(BaseStashQueue):
)
except DvcException:
raise
except Exception as exc: # noqa: BLE001
except Exception as exc:
raise DvcException(f"Failed to reproduce experiment '{rev[:7]}'") from exc
finally:
executor.cleanup(infofile)
+3 -3
View File
@@ -44,9 +44,9 @@ def remove( # noqa: C901, PLR0912
exp_ref_list: list["ExpRefInfo"] = []
queue_entry_list: list["QueueEntry"] = []
if exp_names:
results: dict[
str, "ExpRefAndQueueEntry"
] = celery_queue.get_ref_and_entry_by_names(exp_names, git_remote)
results: dict[str, "ExpRefAndQueueEntry"] = (
celery_queue.get_ref_and_entry_by_names(exp_names, git_remote)
)
remained: list[str] = []
for name, result in results.items():
if not result.exp_ref_info and not result.queue_entry:
+1 -1
View File
@@ -306,7 +306,7 @@ def format_time(
return fill_value
if iso:
return timestamp.isoformat()
if timestamp.date() == date.today():
if timestamp.date() == date.today(): # noqa: DTZ011
fmt = "%I:%M %p"
else:
fmt = "%b %d, %Y"
+1 -1
View File
@@ -287,7 +287,7 @@ def gen_random_name():
# fmt: on
# Use custom random generator to make sure that names are random even if
# global random seed is set (common for ML pipelines).
random_generator = random.Random()
random_generator = random.Random() # noqa: S311
adjective = random_generator.choice(ADJECTIVES)
noun = random_generator.choice(NOUNS)
return f"{adjective}-{noun}"
+5 -4
View File
@@ -2,6 +2,7 @@ from typing import TYPE_CHECKING
from dvc.exceptions import DownloadError
from dvc.log import logger
from dvc.stage.cache import RunCacheNotSupported
from dvc.ui import ui
from dvc_data.index import DataIndex, FileStorage
@@ -49,14 +50,12 @@ def _collect_indexes( # noqa: PLR0913
config["core"] = core
def stage_filter(stage: "Stage") -> bool:
if push and stage.is_repo_import:
return False
return True
return not (push and stage.is_repo_import)
def outs_filter(out: "Output") -> bool:
if push and not out.can_push:
return False
return True
return not (remote and out.remote and remote != out.remote)
for rev in repo.brancher(
revs=revs,
@@ -138,6 +137,8 @@ def fetch( # noqa: PLR0913
try:
if run_cache:
self.stage_cache.pull(remote)
except RunCacheNotSupported as e:
logger.debug("failed to pull run cache: %s", e)
except DownloadError as exc:
failed_count += exc.amount
+1 -1
View File
@@ -50,7 +50,7 @@ def _used_obj_ids_not_in_remote(
@locked
def gc( # noqa: PLR0913, C901
def gc( # noqa: C901, PLR0912, PLR0913
self: "Repo",
all_branches: bool = False,
cloud: bool = False,
+20 -3
View File
@@ -1,3 +1,4 @@
import logging
import os
from collections.abc import Iterable, Iterator
from itertools import chain
@@ -8,7 +9,7 @@ from scmrepo.exceptions import SCMError
from dvc.log import logger
from dvc.scm import NoSCMError
from dvc.utils import as_posix, expand_paths
from dvc.utils import as_posix
from dvc.utils.collections import ensure_list
from dvc.utils.serialize import load_path
@@ -103,7 +104,7 @@ def _collect_metrics(
paths = (fs.from_os_path(metric) for metric in metrics)
# make paths absolute for DVCFileSystem
repo_paths = (f"{fs.root_marker}{path}" for path in paths)
return ldistinct(expand_paths(fs, repo_paths))
return ldistinct(try_expand_paths(fs, repo_paths))
class FileResult(TypedDict, total=False):
@@ -116,6 +117,22 @@ class Result(TypedDict, total=False):
error: Exception
def try_expand_paths(fs: "FileSystem", paths: Iterable[str]) -> Iterator[str]:
for path in paths:
try:
if fs.isdir(path):
yield from fs.find(path)
continue
except Exception as e: # noqa: BLE001
logger.debug(
"failed to expand %r: %s",
path,
e,
exc_info=logger.isEnabledFor(logging.TRACE), # type: ignore[attr-defined]
)
yield path
def to_relpath(fs: "FileSystem", root_dir: str, d: Result) -> Result:
relpath = fs.relpath
cwd = fs.getcwd()
@@ -207,7 +224,7 @@ def show(
on_error=on_error,
)
res[rev] = Result(data=result)
except Exception as exc: # noqa: BLE001
except Exception as exc:
if on_error == "raise":
raise
+4 -4
View File
@@ -6,9 +6,9 @@ from typing import TYPE_CHECKING, Any, Optional, Union
from dvc.dependency.param import ParamsDependency, read_param_file
from dvc.log import logger
from dvc.repo.metrics.show import FileResult, Result
from dvc.repo.metrics.show import FileResult, Result, try_expand_paths
from dvc.stage import PipelineStage
from dvc.utils import as_posix, expand_paths
from dvc.utils import as_posix
from dvc.utils.collections import ensure_list
if TYPE_CHECKING:
@@ -77,7 +77,7 @@ def _collect_params(
path = fs.from_os_path(param)
# make paths absolute for DVCFileSystem
repo_path = f"{fs.root_marker}{path}"
ret.update(dict.fromkeys(expand_paths(fs, [repo_path]), _params))
ret.update(dict.fromkeys(try_expand_paths(fs, [repo_path]), _params))
return ret
@@ -191,7 +191,7 @@ def show(
on_error=on_error,
)
res[rev] = Result(data=params)
except Exception as exc: # noqa: BLE001
except Exception as exc:
if on_error == "raise":
raise
logger.warning("failed to load params in revision %r, %s", rev, str(exc))
+8 -9
View File
@@ -369,11 +369,9 @@ def _matches(targets, config_file, plot_id):
return True
full_id = get_plot_id(plot_id, config_file)
if any(
return any(
(re.match(target, plot_id) or re.match(target, full_id)) for target in targets
):
return True
return False
)
def _normpath(path):
@@ -462,11 +460,12 @@ def _resolve_definitions(
fs, data_path, props=plot_props | props, onerror=onerror
)
# use config for parent directory with most specific definition
unpacked["data"] = {
k: v
for k, v in unpacked["data"].items()
if _closest_parent(fs, k, plot_ids_parents) == data_path
}
if unpacked.get("data"):
unpacked["data"] = {
k: v
for k, v in unpacked["data"].items()
if _closest_parent(fs, k, plot_ids_parents) == data_path
}
dpath.merge(result, unpacked)
elif _matches(targets, config_path, plot_id):
adjusted_props = _adjust_sources(fs, plot_props, config_dir)
+9 -2
View File
@@ -1,10 +1,14 @@
from contextlib import suppress
from dvc.exceptions import InvalidArgumentError, UploadError
from dvc.log import logger
from dvc.stage.cache import RunCacheNotSupported
from dvc.ui import ui
from . import locked
logger = logger.getChild(__name__)
def _rebuild(idx, path, fs, cb):
from dvc_data.index import DataIndex, DataIndexEntry, Meta
@@ -95,8 +99,11 @@ def push( # noqa: PLR0913
"Multiple rev push is unsupported for cloud versioned remotes"
)
used_run_cache = self.stage_cache.push(remote) if run_cache else []
transferred_count += len(used_run_cache)
try:
used_run_cache = self.stage_cache.push(remote) if run_cache else []
transferred_count += len(used_run_cache)
except RunCacheNotSupported as e:
logger.debug("failed to push run cache: %s", e)
if isinstance(targets, str):
targets = [targets]
+2 -6
View File
@@ -25,15 +25,11 @@ def build_outs_trie(stages):
overlapping = out
if parent and overlapping:
msg = (
"The output paths:\n'{}'('{}')\n'{}'('{}')\n"
f"The output paths:\n'{parent!s}'('{parent.stage.addressing}')\n"
f"'{overlapping!s}'('{overlapping.stage.addressing}')\n"
"overlap and are thus in the same tracked directory.\n"
"To keep reproducibility, outputs should be in separate "
"tracked directories or tracked individually."
).format(
str(parent),
parent.stage.addressing,
str(overlapping),
overlapping.stage.addressing,
)
raise OverlappingOutputPathsError(parent, overlapping, msg)
+2 -6
View File
@@ -76,14 +76,10 @@ def worktree_view(
"""
def stage_filter(stage: "Stage") -> bool:
if push and stage.is_repo_import:
return False
return True
return not (push and stage.is_repo_import)
def outs_filter(out: "Output") -> bool:
if not out.is_in_repo or not out.use_cache or (push and not out.can_push):
return False
return True
return out.is_in_repo and out.use_cache and (not push or out.can_push)
return index.targets_view(
targets,
+7 -7
View File
@@ -1,4 +1,5 @@
"""Manages source control systems (e.g. Git)."""
import os
from collections.abc import Iterator, Mapping
from contextlib import contextmanager
@@ -70,8 +71,7 @@ def SCM(
*,
search_parent_directories: bool = ...,
no_scm: Literal[False] = ...,
) -> "Git":
...
) -> "Git": ...
@overload
@@ -80,8 +80,7 @@ def SCM(
*,
search_parent_directories: bool = ...,
no_scm: Literal[True],
) -> "NoSCM":
...
) -> "NoSCM": ...
@overload
@@ -90,8 +89,7 @@ def SCM(
*,
search_parent_directories: bool = ...,
no_scm: bool = ...,
) -> Union["Git", "NoSCM"]:
...
) -> Union["Git", "NoSCM"]: ...
def SCM(root_dir, *, search_parent_directories=True, no_scm=False):
@@ -241,7 +239,9 @@ def iter_revs(
if commit_date:
from datetime import datetime
commit_datestamp = datetime.strptime(commit_date, "%Y-%m-%d").timestamp()
commit_datestamp = (
datetime.strptime(commit_date, "%Y-%m-%d").timestamp() # noqa: DTZ007
)
def _time_filter(rev):
try:
+19 -14
View File
@@ -339,7 +339,7 @@ class Stage(params.StageParams):
for out in self.outs:
status = out.status()
if status:
if allow_missing and status[str(out)] == "not in cache":
if allow_missing and status[str(out)] in ["not in cache", "deleted"]:
continue
logger.debug(
"Output '%s' of %s changed because it is '%s'.",
@@ -418,17 +418,28 @@ class Stage(params.StageParams):
@rwlocked(read=["deps"], write=["outs"])
def reproduce(self, interactive=False, **kwargs) -> Optional["Stage"]:
if not (
kwargs.get("force", False)
or self.changed(
kwargs.get("allow_missing", False), kwargs.pop("upstream", None)
)
):
force = kwargs.get("force", False)
allow_missing = kwargs.get("allow_missing", False)
pull = kwargs.get("pull", False)
upstream = kwargs.pop("upstream", None)
if force:
pass
# Skip stages with missing data if otherwise unchanged
elif not self.changed(allow_missing, upstream):
if not isinstance(self, PipelineStage) and self.is_data_source:
logger.info("'%s' didn't change, skipping", self.addressing)
else:
logger.info("Stage '%s' didn't change, skipping", self.addressing)
return None
# Pull stages with missing data if otherwise unchanged
elif not self.changed(True, upstream) and pull:
try:
logger.info("Pulling data for %s", self)
self.repo.pull(self.addressing, jobs=kwargs.get("jobs", None))
self.checkout()
return None
except CheckoutError:
logger.info("Unable to pull data for %s", self)
msg = f"Going to reproduce {self}. Are you sure you want to continue?"
if interactive and not prompt.confirm(msg):
@@ -597,16 +608,10 @@ class Stage(params.StageParams):
if (self.cmd or self.is_import) and not self.frozen and not dry:
self.remove_outs(ignore_remove=False, force=False)
if (
self.is_import and (not self.frozen or kwargs.get("pull"))
) or self.is_partial_import:
if (self.is_import and not self.frozen) or self.is_partial_import:
self._sync_import(dry, force, kwargs.get("jobs", None), no_download)
elif not self.frozen and self.cmd:
self._run_stage(dry, force, **kwargs)
elif kwargs.get("pull"):
logger.info("Pulling data for %s", self)
self.repo.pull(self.addressing, jobs=kwargs.get("jobs", None))
self.checkout()
elif not dry:
args = ("outputs", "frozen ") if self.frozen else ("data sources", "")
logger.info("Verifying %s in %s%s", *args, self)
+57 -17
View File
@@ -2,6 +2,9 @@ import os
import signal
import subprocess
import threading
from functools import cache
from packaging.version import InvalidVersion, Version
from dvc.log import logger
from dvc.utils import fix_env
@@ -12,25 +15,59 @@ from .exceptions import StageCmdFailedError
logger = logger.getChild(__name__)
@cache
def _fish_supports_no_config(executable) -> bool:
"""
Check if the fish shell supports the --no-config option.
Parameters:
executable (str): The path to the fish shell executable.
Returns:
bool: True if the fish version is greater than 3.3.0, False otherwise.
"""
try:
output = subprocess.check_output(
[executable, "--version"], # noqa: S603
text=True,
)
version = Version(output.split(" ")[-1].strip())
version_to_check = Version("3.3.0")
return version >= version_to_check
except (subprocess.CalledProcessError, IndexError, InvalidVersion):
logger.trace("could not check fish version, defaulting to False")
return False
def _warn_if_fish(executable):
if (
executable is None
or os.path.basename(executable) != "fish"
or _fish_supports_no_config(executable)
):
return
logger.warning(
"DVC detected that you are using a version of fish shell below 3.3.0 "
"Be aware that it might cause problems by overwriting "
"your current environment variables with values defined "
"in 'config.fish', which might affect your command. See "
"https://github.com/iterative/dvc/issues/1307. "
)
def _make_cmd(executable, cmd):
if executable is None:
return cmd
opts = {"zsh": ["--no-rcs"], "bash": ["--noprofile", "--norc"]}
opts = {
"zsh": ["--no-rcs"],
"bash": ["--noprofile", "--norc"],
"fish": [],
}
name = os.path.basename(executable).lower()
return [executable, *opts.get(name, []), "-c", cmd]
def warn_if_fish(executable):
if executable is None or os.path.basename(os.path.realpath(executable)) != "fish":
return
logger.warning(
"DVC detected that you are using fish as your default "
"shell. Be aware that it might cause problems by overwriting "
"your current environment variables with values defined "
"in '.fishrc', which might affect your command. See "
"https://github.com/iterative/dvc/issues/1307. "
)
opt = opts.get(name, [])
if name == "fish" and _fish_supports_no_config(executable):
opt.append("--no-config")
return [executable, *opt, "-c", cmd]
def _enforce_cmd_list(cmd):
@@ -39,7 +76,7 @@ def _enforce_cmd_list(cmd):
def prepare_kwargs(stage, run_env=None):
from dvc.env import DVC_ROOT
from dvc.env import DVC_ROOT, DVC_STAGE
kwargs = {"cwd": stage.wdir, "env": fix_env(None), "close_fds": True}
@@ -48,6 +85,9 @@ def prepare_kwargs(stage, run_env=None):
if DVC_ROOT not in kwargs["env"]:
kwargs["env"][DVC_ROOT] = stage.repo.root_dir
# Create DVC_STAGE env variable for every command
kwargs["env"][DVC_STAGE] = stage.addressing
# NOTE: when you specify `shell=True`, `Popen` [1] will default to
# `/bin/sh` on *nix and will add ["/bin/sh", "-c"] to your command.
# But we actually want to run the same shell that we are running
@@ -103,7 +143,7 @@ def cmd_run(stage, dry=False, run_env=None):
executable = get_executable()
if not dry:
warn_if_fish(executable)
_warn_if_fish(executable)
for cmd in commands:
display_command(cmd)
+1 -3
View File
@@ -92,9 +92,7 @@ def check_no_externals(stage):
from dvc.utils import format_link
def _is_cached_external(out):
if out.is_in_repo or not out.use_cache:
return False
return True
return not out.is_in_repo and out.use_cache
outs = [str(out) for out in stage.outs if _is_cached_external(out)]
if not outs:
@@ -1,6 +1,7 @@
"""
Tests for modifications to an existing dataset.
"""
import glob
import os
import random
+5
View File
@@ -104,6 +104,11 @@ def make_dvc_bin(
venv.run("pip", "install", f"git+file://{dvc_git_repo}@{dvc_rev}#egg={egg}")
if dvc_rev in ["2.18.1", "2.11.0", "2.6.3"]:
venv.run("pip", "install", "fsspec==2022.11.0")
try:
if version.Version(dvc_rev) < version.Version("3.50.3"):
venv.run("pip", "install", "pygit2==1.14.1")
except version.InvalidVersion:
pass
dvc_venvs[dvc_rev] = venv
dvc_bin = venv.which("dvc")
else:
-11
View File
@@ -5,7 +5,6 @@ import json
import os
import re
import sys
from collections.abc import Iterable, Iterator
from typing import TYPE_CHECKING, Optional
import colorama
@@ -13,8 +12,6 @@ import colorama
if TYPE_CHECKING:
from typing import TextIO
from dvc.fs import FileSystem
LARGE_DIR_SIZE = 100
TARGET_REGEX = re.compile(r"(?P<path>.*?)(:(?P<name>[^\\/:]*))??$")
@@ -410,14 +407,6 @@ def errored_revisions(rev_data: dict) -> list:
return result
def expand_paths(fs: "FileSystem", paths: Iterable[str]) -> Iterator[str]:
for path in paths:
if fs.isdir(path):
yield from fs.find(path)
else:
yield path
def isatty(stream: "Optional[TextIO]") -> bool:
if stream is None:
return False
+1 -3
View File
@@ -39,9 +39,7 @@ def apply_diff(src, dest): # noqa: C901
dest[i] = value
else:
raise AssertionError( # noqa: TRY004
"Can't apply diff from {} to {}".format(
src.__class__.__name__, dest.__class__.__name__
)
f"Can't apply diff from {type(src).__name__} to {type(dest).__name__}"
)
+5 -8
View File
@@ -1,4 +1,5 @@
"""Common utilities for serialize."""
import os
from contextlib import AbstractContextManager, contextmanager
from typing import TYPE_CHECKING, Any, Callable, Optional, Protocol, TextIO, Union
@@ -15,25 +16,21 @@ if TYPE_CHECKING:
class DumperFn(Protocol):
def __call__(
self, path: "StrPath", data: Any, fs: Optional["FileSystem"] = None
) -> Any:
...
) -> Any: ...
class DumpersFn(Protocol):
def __call__(self, data: Any, stream: TextIO) -> Any:
...
def __call__(self, data: Any, stream: TextIO) -> Any: ...
class ModifierFn(Protocol):
def __call__(
self, path: "StrPath", fs: Optional["FileSystem"] = None
) -> AbstractContextManager[dict]:
...
) -> AbstractContextManager[dict]: ...
class LoaderFn(Protocol):
def __call__(self, path: "StrPath", fs: Optional["FileSystem"] = None) -> Any:
...
def __call__(self, path: "StrPath", fs: Optional["FileSystem"] = None) -> Any: ...
ReadType = Union[bytes, None, str]
+1
View File
@@ -6,6 +6,7 @@ Used for parsing dvc.yaml, dvc.lock and .dvc files.
Not to be confused with strictyaml, a python library with similar motivations.
"""
import re
import typing
from contextlib import suppress
+17
View File
@@ -1,3 +1,4 @@
import os
from typing import TYPE_CHECKING, Any, Optional
from urllib.parse import urljoin
@@ -125,3 +126,19 @@ def get_subrepo_relpath(repo: "Repo") -> str:
relpath = as_posix(repo.fs.relpath(repo.root_dir, scm_root_dir))
return "" if relpath == "." else relpath
def get_repo_url(repo: "Repo") -> str:
from dulwich.porcelain import get_remote_repo
from dvc.env import DVC_EXP_GIT_REMOTE
repo_url = os.getenv(
DVC_EXP_GIT_REMOTE, repo.config.get("exp", {}).get("git_remote")
)
if repo_url:
try:
_, repo_url = get_remote_repo(repo.scm.dulwich.repo, repo_url)
except IndexError:
pass
return repo_url
+15 -9
View File
@@ -31,24 +31,30 @@ classifiers = [
dynamic = ["version"]
dependencies = [
"attrs>=22.2.0",
"celery",
"colorama>=0.3.9",
"configobj>=5.0.6",
"distro>=1.3",
"dpath<3,>=2.1.0",
"dvc-data>=3.13,<3.15",
"dulwich",
"dvc-data>=3.15,<3.16",
"dvc-http>=2.29.0",
"dvc-objects",
"dvc-render>=1.0.1,<2",
"dvc-studio-client>=0.20,<1",
"dvc-task>=0.3.0,<1",
"flatten_dict<1,>=0.4.1",
# https://github.com/iterative/dvc/issues/9654
"flufl.lock>=5,<8",
"flufl.lock>=5,<8", # https://github.com/iterative/dvc/issues/9654
"fsspec",
"funcy>=1.14",
"grandalf<1,>=0.7",
"gto>=1.6.0,<2",
"hydra-core>=1.1",
"iterative-telemetry>=0.0.7",
"kombu",
"networkx>=2.5",
"omegaconf",
"packaging>=19",
"pathspec>=0.10.3",
"platformdirs<4,>=3.1.1",
@@ -59,7 +65,7 @@ dependencies = [
"requests>=2.22",
"rich>=12",
"ruamel.yaml>=0.17.11",
"scmrepo>=3,<4",
"scmrepo>=3.3.4,<4",
"shortuuid>=0.5",
"shtab<2,>=1.3.4",
"tabulate>=0.8.7",
@@ -77,7 +83,7 @@ gdrive = ["dvc-gdrive>=3,<4"]
gs = ["dvc-gs>=3,<4"]
hdfs = ["dvc-hdfs>=3,<4"]
lint = [
"mypy==1.8.0",
"mypy==1.10.0",
"pandas-stubs",
"types-colorama",
"types-psutil",
@@ -89,7 +95,7 @@ lint = [
"typing-extensions",
]
oss = ["dvc-oss>=3,<4"]
s3 = ["dvc-s3>=3,<4"]
s3 = ["dvc-s3>=3.2.0,<4"]
ssh = ["dvc-ssh>=4,<5"]
ssh_gssapi = ["dvc-ssh[gssapi]>=4,<5"]
testing = [
@@ -261,10 +267,10 @@ ignore = [
"TRY003", "TRY300", "PLR2004", "PLW2901", "LOG007",
]
select = [
"F", "E", "W", "C90", "I", "N", "UP", "YTT", "ASYNC", "S", "BLE", "B", "A", "C4", "T10",
"F", "E", "W", "C90", "I", "N", "UP", "YTT", "ASYNC", "S", "BLE", "B", "A", "C4", "DTZ", "T10",
"EXE", "ISC", "ICN", "G", "INP", "PIE", "T20", "PYI", "PT", "Q", "RSE", "RET",
"SLOT", "SIM", "TID", "TCH", "ARG", "PGH", "PLC", "PLE", "PLR", "PLW", "TRY",
"FLY", "PERF101", "LOG", "RUF", "RUF022", "RUF023", "RUF024", "RUF025", "RUF026",
"FLY", "PERF101", "LOG", "RUF", "RUF022", "RUF023", "RUF024", "RUF025", "RUF026", "RUF027"
]
preview = true
explicit-preview-rules = true
@@ -286,7 +292,7 @@ strict = true
ignore-variadic-names = true
[tool.ruff.lint.isort]
known-first-party = ["dvc", "dvc_data", "dvc_objects", "dvc_render", "dvc_task", "tests"]
known-first-party = ["dvc", "dvc_*", "tests"]
[tool.ruff.lint.pep8-naming]
extend-ignore-names = ["M", "SCM"]
-1
View File
@@ -43,7 +43,6 @@ from global repo template to creating everything inplace, which:
- does not create unnecessary files
"""
import os
import pytest
+74
View File
@@ -0,0 +1,74 @@
from os.path import join, normpath
import pytest
from dvc.api import artifacts_show
from dvc.testing.tmp_dir import make_subrepo
from dvc.utils import as_posix
from tests.func.artifacts.test_artifacts import get_tag_and_name, make_artifact
@pytest.mark.parametrize("sub", ["sub", ""])
def test_artifacts_show(tmp_dir, dvc, scm, sub):
subdir = tmp_dir / sub
dirname = str(subdir.relative_to(tmp_dir))
tag, name = get_tag_and_name(as_posix(dirname), "myart", "v2.0.0")
make_artifact(tmp_dir, "myart", tag, subdir / "myart.pkl")
assert artifacts_show(name) == {
"path": normpath(join(dirname, "myart.pkl")),
"rev": scm.get_rev(),
}
assert artifacts_show(name, repo=tmp_dir.fs_path) == {
"path": normpath(join(dirname, "myart.pkl")),
"rev": scm.get_rev(),
}
assert artifacts_show(name, repo=f"file://{tmp_dir.as_posix()}") == {
"path": normpath(join(dirname, "myart.pkl")),
"rev": scm.get_rev(),
}
assert artifacts_show(name, repo=subdir.fs_path) == {
"path": normpath(join(dirname, "myart.pkl")),
"rev": scm.get_rev(),
}
with subdir.chdir():
assert artifacts_show(name) == {
"path": normpath(join(dirname, "myart.pkl")),
"rev": scm.get_rev(),
}
@pytest.mark.parametrize("sub", ["sub", ""])
def test_artifacts_show_subrepo(tmp_dir, scm, sub):
subrepo = tmp_dir / "subrepo"
make_subrepo(subrepo, scm)
subdir = subrepo / sub
dirname = str(subdir.relative_to(tmp_dir))
tag, name = get_tag_and_name(as_posix(dirname), "myart", "v2.0.0")
make_artifact(subrepo, "myart", tag, subdir / "myart.pkl")
assert artifacts_show(name) == {
"path": join(dirname, "myart.pkl"),
"rev": scm.get_rev(),
}
assert artifacts_show(name, repo=tmp_dir.fs_path) == {
"path": join(dirname, "myart.pkl"),
"rev": scm.get_rev(),
}
assert artifacts_show(name, repo=f"file://{tmp_dir.as_posix()}") == {
"path": join(dirname, "myart.pkl"),
"rev": scm.get_rev(),
}
assert artifacts_show(name, repo=subdir.fs_path) == {
"path": str((subdir / "myart.pkl").relative_to(subrepo)),
"rev": scm.get_rev(),
}
with subdir.chdir():
assert artifacts_show(name) == {
"path": str((subdir / "myart.pkl").relative_to(subrepo)),
"rev": scm.get_rev(),
}
+43
View File
@@ -31,6 +31,40 @@ def test_get_url_requires_dvc(tmp_dir, scm):
api.get_url("foo", repo=f"file://{tmp_dir.as_posix()}")
def test_get_url_from_remote(tmp_dir, erepo_dir, cloud, local_cloud):
erepo_dir.add_remote(config=cloud.config, name="other")
erepo_dir.add_remote(config=local_cloud.config, default=True)
with erepo_dir.chdir():
erepo_dir.dvc_gen("foo", "foo", commit="add foo")
# Using file url to force clone to tmp repo
repo_url = f"file://{erepo_dir.as_posix()}"
expected_rel_path = os.path.join(
"files", "md5", "ac/bd18db4cc2f85cedef654fccc4a4d8"
)
# Test default remote
assert api.get_url("foo", repo=repo_url) == (local_cloud / expected_rel_path).url
# Test remote arg
assert (
api.get_url("foo", repo=repo_url, remote="other")
== (cloud / expected_rel_path).url
)
# Test config arg
assert (
api.get_url("foo", repo=repo_url, config={"core": {"remote": "other"}})
== (cloud / expected_rel_path).url
)
# Test remote_config arg
assert (
api.get_url("foo", repo=repo_url, remote_config={"url": cloud.url})
== (cloud / expected_rel_path).url
)
def test_open_external(tmp_dir, erepo_dir, cloud):
erepo_dir.add_remote(config=cloud.config)
@@ -253,3 +287,12 @@ def test_read_from_remote(tmp_dir, erepo_dir, cloud, local_cloud):
)
== "foo content"
)
assert (
api.read(
os.path.join("dir", "foo"),
repo=f"file://{erepo_dir.as_posix()}",
remote_config={"url": cloud.url},
)
== "foo content"
)
+96 -2
View File
@@ -6,7 +6,9 @@ import pytest
from dvc.annotations import Artifact
from dvc.exceptions import ArtifactNotFoundError, InvalidArgumentError
from dvc.repo.artifacts import check_name_format
from dvc.repo.artifacts import Artifacts, check_name_format
from dvc.testing.tmp_dir import make_subrepo
from dvc.utils import as_posix
from dvc.utils.strictyaml import YAMLSyntaxError, YAMLValidationError
dvcyaml = {
@@ -185,7 +187,7 @@ def test_get_rev(tmp_dir, dvc, scm):
dvc.artifacts.get_rev("myart", stage="prod")
def test_get_path(tmp_dir, dvc):
def test_get_path(tmp_dir, dvc, scm):
(tmp_dir / "dvc.yaml").dump(dvcyaml)
subdir = tmp_dir / "subdir"
subdir.mkdir()
@@ -196,3 +198,95 @@ def test_get_path(tmp_dir, dvc):
assert dvc.artifacts.get_path("subdir/dvc.yaml:myart") == os.path.join(
"subdir", "myart.pkl"
)
def test_parametrized(tmp_dir, dvc):
(tmp_dir / "params.yaml").dump({"path": "myart.pkl"})
(tmp_dir / "dvc.yaml").dump(
{"artifacts": {"myart": {"type": "model", "path": "${path}"}}}
)
assert tmp_dir.dvc.artifacts.read() == {
"dvc.yaml": {"myart": Artifact(path="myart.pkl", type="model")}
}
def test_get_path_subrepo(tmp_dir, scm, dvc):
subrepo = tmp_dir / "subrepo"
make_subrepo(subrepo, scm)
(subrepo / "dvc.yaml").dump(dvcyaml)
assert dvc.artifacts.get_path("subrepo:myart") == os.path.join(
"subrepo", "myart.pkl"
)
assert dvc.artifacts.get_path("subrepo/dvc.yaml:myart") == os.path.join(
"subrepo", "myart.pkl"
)
assert subrepo.dvc.artifacts.get_path("subrepo:myart") == os.path.join(
"subrepo", "myart.pkl"
)
assert subrepo.dvc.artifacts.get_path("subrepo/dvc.yaml:myart") == os.path.join(
"subrepo", "myart.pkl"
)
def get_tag_and_name(dirname, name, version):
tagname = f"{name}@{version}"
if dirname in (os.curdir, ""):
return tagname, name
return f"{dirname}={tagname}", f"{dirname}:{name}"
def make_artifact(tmp_dir, name, tag, path) -> Artifact:
artifact = Artifact(path=path.name, type="model")
dvcfile = path.with_name("dvc.yaml")
tmp_dir.scm_gen(path, "hello_world", commit="add myart.pkl")
tmp_dir.dvc.artifacts.add(name, artifact, dvcfile=os.fspath(dvcfile))
tmp_dir.scm.add_commit([dvcfile], message="add dvc.yaml")
tmp_dir.scm.tag(tag, annotated=True, message="foo")
return artifact
@pytest.mark.parametrize("sub", ["sub", ""])
def test_artifacts_download(tmp_dir, dvc, scm, sub):
subdir = tmp_dir / sub
dirname = str(subdir.relative_to(tmp_dir))
tag, name = get_tag_and_name(as_posix(dirname), "myart", "v2.0.0")
make_artifact(tmp_dir, "myart", tag, subdir / "myart.pkl")
result = (1, "myart.pkl")
assert Artifacts.get(".", name, force=True) == result
assert Artifacts.get(tmp_dir.fs_path, name, force=True) == result
assert Artifacts.get(f"file://{tmp_dir.as_posix()}", name, force=True) == result
assert Artifacts.get(subdir.fs_path, name, force=True) == result
with subdir.chdir():
assert Artifacts.get(".", name, force=True) == result
@pytest.mark.parametrize("sub", ["sub", ""])
def test_artifacts_download_subrepo(tmp_dir, scm, sub):
subrepo = tmp_dir / "subrepo"
make_subrepo(subrepo, scm)
subdir = subrepo / sub
dirname = str(subdir.relative_to(tmp_dir))
tag, name = get_tag_and_name(as_posix(dirname), "myart", "v2.0.0")
make_artifact(subrepo, "myart", tag, subdir / "myart.pkl")
result = (1, "myart.pkl")
assert Artifacts.get(".", name) == result
assert Artifacts.get(tmp_dir.fs_path, name, force=True) == result
assert Artifacts.get(f"file://{tmp_dir.as_posix()}", name, force=True) == result
assert Artifacts.get(subdir.fs_path, name, force=True) == result
with subdir.chdir():
assert Artifacts.get(".", name, force=True) == result
def test_artifacts_download_studio(tmp_dir, dvc, mocker):
with dvc.config.edit("global") as conf:
conf["studio"]["token"] = "mytoken"
download_studio = mocker.patch("dvc.repo.artifacts.Artifacts._download_studio")
Artifacts.get("myart.pkl", "myart.pkl")
assert download_studio.call_args.kwargs["dvc_studio_config"]["token"] == "mytoken"
+28
View File
@@ -409,3 +409,31 @@ def test_auto_push_on_save(
dvc.experiments.save(name=exp_name, force=True)
assert first(dvc.experiments.push(name=exp_name, git_remote=remote)) == expected_key
def test_auto_push_misconfigured(
tmp_dir, scm, dvc, git_upstream, local_remote, exp_stage, caplog
):
with dvc.config.edit() as conf:
conf["exp"]["auto_push"] = True
conf["exp"]["git_remote"] = "notfound"
exp_name = "foo"
with caplog.at_level(logging.WARNING, logger="dvc"):
dvc.experiments.run(exp_stage.addressing, params=["foo=2"], name=exp_name)
assert "Failed to validate remotes" in caplog.text
def test_auto_push_tmp_dir(tmp_dir, scm, dvc, git_upstream, local_remote, exp_stage):
remote = git_upstream.remote
with dvc.config.edit() as conf:
conf["exp"]["auto_push"] = True
conf["exp"]["git_remote"] = remote
exp_name = "foo"
dvc.experiments.run(
exp_stage.addressing, params=["foo=2"], name=exp_name, tmp_dir=True
)
assert first(dvc.experiments.push(name=exp_name, git_remote=remote)) == "up_to_date"
+5 -9
View File
@@ -101,7 +101,7 @@ def test_show_simple(tmp_dir, scm, dvc, exp_stage):
@pytest.mark.parametrize("workspace", [True, False])
def test_show_experiment(tmp_dir, scm, dvc, exp_stage, workspace):
baseline_rev = scm.get_rev()
timestamp = datetime.fromtimestamp(
timestamp = datetime.fromtimestamp( # noqa: DTZ006
scm.gitpython.repo.rev_parse(baseline_rev).committed_date
)
@@ -423,7 +423,7 @@ def test_show_csv(tmp_dir, scm, dvc, exp_stage, capsys):
baseline_rev = scm.get_rev()
def _get_rev_isotimestamp(rev):
return datetime.fromtimestamp(
return datetime.fromtimestamp( # noqa: DTZ006
scm.gitpython.repo.rev_parse(rev).committed_date
).isoformat()
@@ -453,15 +453,11 @@ def test_show_csv(tmp_dir, scm, dvc, exp_stage, capsys):
in cap.out
)
assert (
"{},{},branch_base,{},,2,2,{}".format(
ref_info1.name, rev1[:7], _get_rev_isotimestamp(rev1), data_hash
)
f"{ref_info1.name},{rev1[:7]},branch_base,{_get_rev_isotimestamp(rev1)},,2,2,{data_hash}"
in cap.out
)
assert (
"{},{},branch_commit,{},,3,3,{}".format(
ref_info2.name, rev2[:7], _get_rev_isotimestamp(rev2), data_hash
)
f"{ref_info2.name},{rev2[:7]},branch_commit,{_get_rev_isotimestamp(rev2)},,3,3,{data_hash}"
in cap.out
)
@@ -656,7 +652,7 @@ def test_metrics_renaming(tmp_dir, dvc, scm, capsys, copy_script):
cap = capsys.readouterr()
def _get_rev_isotimestamp(rev):
return datetime.fromtimestamp(
return datetime.fromtimestamp( # noqa: DTZ006
scm.gitpython.repo.rev_parse(rev).committed_date
).isoformat()
+27
View File
@@ -1,5 +1,6 @@
import json
import os
import shutil
from os.path import join
import pytest
@@ -9,8 +10,10 @@ from dvc.cli import main
from dvc.dvcfile import PROJECT_FILE
from dvc.exceptions import OverlappingOutputPathsError
from dvc.repo import Repo
from dvc.repo.metrics.show import FileResult, Result
from dvc.utils.fs import remove
from dvc.utils.serialize import JSONFileCorruptedError
from dvc_data.index import DataIndexDirError
def test_show_simple(tmp_dir, dvc, run_copy_metrics):
@@ -342,3 +345,27 @@ def test_cached_metrics(tmp_dir, dvc, scm, remote):
}
}
}
def test_top_level_parametrized(tmp_dir, dvc):
tmp_dir.dvc_gen("metrics.yaml", "foo: 3\nbar: 10")
(tmp_dir / "params.yaml").dump({"metric_file": "metrics.yaml"})
(tmp_dir / "dvc.yaml").dump({"metrics": ["${metric_file}"]})
assert dvc.metrics.show() == {
"": {"data": {"metrics.yaml": {"data": {"foo": 3, "bar": 10}}}}
}
def test_metric_in_a_tracked_directory_with_missing_dir_file(M, tmp_dir, dvc):
tmp_dir.dvc_gen({"dir": {"file": "2"}})
(tmp_dir / "dvc.yaml").dump({"metrics": [join("dir", "file")]})
shutil.rmtree(tmp_dir / "dir") # remove from workspace
dvc.cache.local.clear() # remove .dir file
assert dvc.metrics.show() == {
"": Result(
data={
join("dir", "file"): FileResult(error=M.instance_of(DataIndexDirError)),
}
)
}
+33 -1
View File
@@ -1,9 +1,12 @@
import shutil
from os.path import join
import pytest
from dvc.dvcfile import PROJECT_FILE
from dvc.repo import Repo
from dvc.repo.stage import PROJECT_FILE
from dvc.repo.metrics.show import FileResult, Result
from dvc_data.index import DataIndexDirError
def test_show_empty(dvc):
@@ -178,3 +181,32 @@ def test_cached_params(tmp_dir, dvc, scm, remote):
}
}
}
def test_top_level_parametrized(tmp_dir, dvc):
(tmp_dir / "param.json").dump({"foo": 3, "bar": 10})
(tmp_dir / "params.yaml").dump({"param_file": "param.json"})
(tmp_dir / "dvc.yaml").dump({"params": ["${param_file}"]})
assert dvc.params.show() == {
"": {
"data": {
"param.json": {"data": {"foo": 3, "bar": 10}},
"params.yaml": {"data": {"param_file": "param.json"}},
}
}
}
def test_param_in_a_tracked_directory_with_missing_dir_file(M, tmp_dir, dvc):
tmp_dir.dvc_gen({"dir": {"file": "2"}})
(tmp_dir / "dvc.yaml").dump({"params": [join("dir", "file")]})
shutil.rmtree(tmp_dir / "dir") # remove from workspace
dvc.cache.local.clear() # remove .dir file
assert dvc.params.show() == {
"": Result(
data={
join("dir", "file"): FileResult(error=M.instance_of(DataIndexDirError)),
}
)
}
-1
View File
@@ -1,6 +1,5 @@
"""Negative tests for the parametrization."""
import logging
import re
+1
View File
@@ -1,4 +1,5 @@
"""Testing happy paths for the foreach."""
import os
import pytest
+168
View File
@@ -0,0 +1,168 @@
from dvc.parsing import DataResolver
def test_params(tmp_dir, dvc):
(tmp_dir / "params.yaml").dump(
{"params": {"param1": "params.json", "param2": "params.toml"}}
)
template = {"params": ["${params.param1}", "param11", "${params.param2}"]}
resolver = DataResolver(dvc, tmp_dir, template)
assert resolver.resolve_params() == ["params.json", "param11", "params.toml"]
def test_metrics(tmp_dir, dvc):
(tmp_dir / "params.yaml").dump(
{"metrics": {"metric1": "metrics.json", "metric2": "metrics.toml"}}
)
template = {"metrics": ["${metrics.metric1}", "metric11", "${metrics.metric2}"]}
resolver = DataResolver(dvc, tmp_dir, template)
assert resolver.resolve_metrics() == ["metrics.json", "metric11", "metrics.toml"]
def test_plots(tmp_dir, dvc):
template = {
"plots": [
{
"${plots.plot1_name}": {
"x": "${plots.x_cls}",
"y": {
"train_classes.csv": "${plots.y_train_cls}",
"test_classes.csv": [
"${plots.y_train_cls}",
"${plots.y_test_cls}",
],
},
"title": "Compare test vs train confusion matrix",
"template": "confusion",
"x_label": "Actual class",
"y_label": "Predicted class",
}
},
{"eval/importance2.png": None},
{"${plots.plot3_name}": None},
"eval/importance4.png",
"${plots.plot5_name}",
],
}
(tmp_dir / "params.yaml").dump(
{
"plots": {
"x_cls": "actual_class",
"y_train_cls": "predicted_class",
"y_test_cls": "predicted_class2",
"plot1_name": "eval/importance1.png",
"plot3_name": "eval/importance3.png",
"plot5_name": "eval/importance5.png",
}
}
)
resolver = DataResolver(dvc, tmp_dir, template)
assert resolver.resolve_plots() == [
{
"eval/importance1.png": {
"x": "actual_class",
"y": {
"train_classes.csv": "predicted_class",
"test_classes.csv": ["predicted_class", "predicted_class2"],
},
"title": "Compare test vs train confusion matrix",
"template": "confusion",
"x_label": "Actual class",
"y_label": "Predicted class",
}
},
{"eval/importance2.png": None},
{"eval/importance3.png": None},
"eval/importance4.png",
"eval/importance5.png",
]
def test_artifacts(tmp_dir, dvc):
template = {
"artifacts": {
"${artifacts.name}": {
"path": "${artifacts.path}",
"type": "model",
"desc": "CV classification model, ResNet50",
"labels": ["${artifacts.label1}", "${artifacts.label2}"],
"meta": {"framework": "${artifacts.framework}"},
}
}
}
(tmp_dir / "params.yaml").dump(
{
"artifacts": {
"name": "cv-classification",
"path": "models/resnet.pt",
"label1": "resnet50",
"label2": "classification",
"framework": "pytorch",
}
}
)
resolver = DataResolver(dvc, tmp_dir, template)
assert resolver.resolve_artifacts() == {
"cv-classification": {
"path": "models/resnet.pt",
"type": "model",
"desc": "CV classification model, ResNet50",
"labels": ["resnet50", "classification"],
"meta": {"framework": "pytorch"},
}
}
def test_datasets(tmp_dir, dvc):
template = {
"datasets": [
{"name": "${ds1.name}", "url": "${ds1.url}", "type": "dvcx"},
{
"name": "${ds2.name}",
"url": "${ds2.url}",
"type": "dvc",
"path": "${ds2.path}",
},
{
"name": "${ds3.name}",
"url": "${ds3.url}",
"type": "url",
},
]
}
(tmp_dir / "params.yaml").dump(
{
"ds1": {"name": "dogs", "url": "dvcx://dogs"},
"ds2": {
"name": "example-get-started",
"url": "[email protected]:iterative/example-get-started.git",
"path": "path",
},
"ds3": {
"name": "cloud-versioning-demo",
"url": "s3://cloud-versioning-demo",
},
}
)
resolver = DataResolver(dvc, tmp_dir, template)
assert resolver.resolve_datasets() == [
{"name": "dogs", "url": "dvcx://dogs", "type": "dvcx"},
{
"name": "example-get-started",
"url": "[email protected]:iterative/example-get-started.git",
"type": "dvc",
"path": "path",
},
{
"name": "cloud-versioning-demo",
"url": "s3://cloud-versioning-demo",
"type": "url",
},
]
+70
View File
@@ -500,3 +500,73 @@ def test_show_plots_defined_with_native_os_path(tmp_dir, dvc, scm, capsys):
json_data = json_out["data"]
assert json_data[f"{top_level_plot}"]
assert json_data[stage_plot]
@pytest.mark.parametrize(
"plot_config,expanded_config,expected_datafiles",
[
(
{
"comparison": {
"x": {"${data1}": "${a}"},
"y": {"sub/dir/data2.json": "${b}"},
}
},
{
"comparison": {
"x": {"data1.json": "a"},
"y": {"sub/dir/data2.json": "b"},
}
},
["data1.json", os.path.join("sub", "dir", "data2.json")],
),
(
{"${data1}": None},
{"data1.json": {}},
["data1.json"],
),
(
"${data1}",
{"data1.json": {}},
["data1.json"],
),
],
)
def test_top_level_parametrized(
tmp_dir, dvc, plot_config, expanded_config, expected_datafiles
):
(tmp_dir / "params.yaml").dump(
{"data1": "data1.json", "a": "a", "b": "b", "c": "c"}
)
data = {
"data1.json": [
{"a": 1, "b": 0.1, "c": 0.01},
{"a": 2, "b": 0.2, "c": 0.02},
],
os.path.join("sub", "dir", "data.json"): [
{"a": 6, "b": 0.6, "c": 0.06},
{"a": 7, "b": 0.7, "c": 0.07},
],
}
for filename, content in data.items():
dirname = os.path.dirname(filename)
if dirname:
os.makedirs(dirname)
(tmp_dir / filename).dump_json(content, sort_keys=True)
config_file = "dvc.yaml"
with modify_yaml(config_file) as dvcfile_content:
dvcfile_content["plots"] = [plot_config]
result = dvc.plots.show()
assert expanded_config == get_plot(
result, "workspace", typ="definitions", file=config_file
)
for filename, content in data.items():
if filename in expected_datafiles:
assert content == get_plot(result, "workspace", file=filename)
else:
assert filename not in get_plot(result, "workspace")
@@ -55,3 +55,16 @@ def test_repro_allow_missing_upstream_stage_modified(
ret = dvc.reproduce(pull=True, allow_missing=True)
# create-foo is skipped ; copy-foo pulls modified dep
assert len(ret) == 1
def test_repro_allow_missing_cached(tmp_dir, dvc):
tmp_dir.gen("fixed", "fixed")
dvc.stage.add(name="create-foo", cmd="echo foo > foo", deps=["fixed"], outs=["foo"])
dvc.stage.add(name="copy-foo", cmd="cp foo bar", deps=["foo"], outs=["bar"])
dvc.reproduce()
remove("foo")
ret = dvc.reproduce(allow_missing=True)
# both stages are skipped
assert not ret
+50 -2
View File
@@ -1,5 +1,7 @@
import os
import pytest
from dvc.stage.cache import RunCacheNotSupported
from dvc.utils.fs import remove
@@ -49,11 +51,14 @@ def test_repro_skip_pull_if_no_run_cache_is_passed(tmp_dir, dvc, mocker, local_r
(foo,) = tmp_dir.dvc_gen("foo", "foo")
dvc.push()
spy_pull = mocker.spy(dvc.stage_cache, "pull")
dvc.stage.add(name="copy-foo", cmd="cp foo bar", deps=["foo"], outs=["bar"])
assert dvc.reproduce(pull=True)
remove("foo")
remove(foo.outs[0].cache_path)
remove("dvc.lock")
spy_pull = mocker.spy(dvc.stage_cache, "pull")
assert dvc.reproduce(pull=True, run_cache=False)
assert not spy_pull.called
@@ -68,3 +73,46 @@ def test_repro_skip_pull_if_single_item_is_passed(tmp_dir, dvc, mocker, local_re
remove(foo.outs[0].cache_path)
assert dvc.reproduce(pull=True, single_item=True)
def test_repro_pulls_persisted_output(tmp_dir, dvc, mocker, local_remote):
tmp_dir.dvc_gen("foo", "foo")
dvc.push()
dvc.stage.add(name="copy-foo", cmd="cp foo bar", deps=["foo"], outs_persist=["bar"])
dvc.reproduce()
remove("bar")
# stage is skipped
assert not dvc.reproduce(pull=True)
@pytest.mark.parametrize("allow_missing", [True, False])
def test_repro_pulls_allow_missing(tmp_dir, dvc, mocker, local_remote, allow_missing):
tmp_dir.dvc_gen("foo", "foo")
dvc.push()
dvc.stage.add(name="copy-foo", cmd="cp foo bar", deps=["foo"], outs=["bar"])
dvc.reproduce()
remove("foo")
# stage is skipped
assert not dvc.reproduce(pull=True, allow_missing=allow_missing)
# data only pulled if allow_missing is false
assert (tmp_dir / "foo").exists() != allow_missing
def test_repro_pull_fails(tmp_dir, dvc, mocker, local_remote):
tmp_dir.dvc_gen("foo", "foo")
dvc.stage.add(
name="concat-foo", cmd="cat foo foo > bar", deps=["foo"], outs=["bar"]
)
stages = dvc.reproduce()
remove("bar")
remove(stages[0].outs[0].cache_path)
# stage is run
assert dvc.reproduce(pull=True)
+86
View File
@@ -1,6 +1,7 @@
import logging
import os
import shutil
from os.path import join
import pytest
@@ -8,6 +9,7 @@ import dvc_data
from dvc.cli import main
from dvc.exceptions import CheckoutError
from dvc.repo.open_repo import clean_repos
from dvc.scm import CloneError
from dvc.stage.exceptions import StageNotFound
from dvc.testing.remote_tests import TestRemote # noqa: F401
from dvc.utils.fs import remove
@@ -580,6 +582,58 @@ def test_target_remote(tmp_dir, dvc, make_remote):
}
def test_output_target_remote(tmp_dir, dvc, make_remote):
make_remote("default", default=True)
make_remote("for_foo", default=False)
make_remote("for_bar", default=False)
tmp_dir.dvc_gen("foo", "foo")
tmp_dir.dvc_gen("bar", "bar")
tmp_dir.dvc_gen("data", {"one": "one", "two": "two"})
with (tmp_dir / "foo.dvc").modify() as d:
d["outs"][0]["remote"] = "for_foo"
with (tmp_dir / "bar.dvc").modify() as d:
d["outs"][0]["remote"] = "for_bar"
# push foo and data to for_foo remote
dvc.push(remote="for_foo")
default = dvc.cloud.get_remote_odb("default")
for_foo = dvc.cloud.get_remote_odb("for_foo")
for_bar = dvc.cloud.get_remote_odb("for_bar")
# hashes for foo and data, but not bar
expected = {
"acbd18db4cc2f85cedef654fccc4a4d8",
"f97c5d29941bfb1b2fdab0874906ab82",
"6b18131dc289fd37006705affe961ef8.dir",
"b8a9f715dbb64fd5c56e7783c6820a61",
}
assert set(default.all()) == set()
assert set(for_foo.all()) == expected
assert set(for_bar.all()) == set()
# push everything without specifying remote
dvc.push()
assert set(default.all()) == {
"f97c5d29941bfb1b2fdab0874906ab82",
"6b18131dc289fd37006705affe961ef8.dir",
"b8a9f715dbb64fd5c56e7783c6820a61",
}
assert set(for_foo.all()) == expected
assert set(for_bar.all()) == {"37b51d194a7513e45b56f6524f2d51f2"}
clean(["foo", "bar", "data"], dvc)
# pull foo and data from for_foo remote
dvc.pull(remote="for_foo", allow_missing=True)
assert set(dvc.cache.local.all()) == expected
def test_pull_allow_missing(tmp_dir, dvc, local_remote):
dvc.stage.add(name="bar", outs=["bar"], cmd="echo bar > bar")
@@ -592,3 +646,35 @@ def test_pull_allow_missing(tmp_dir, dvc, local_remote):
stats = dvc.pull(allow_missing=True)
assert stats["fetched"] == 1
def test_pull_granular_excluding_import_that_cannot_be_pulled(
tmp_dir, dvc, local_remote, mocker
):
"""Regression test for https://github.com/iterative/dvc/issues/10309."""
mocker.patch("dvc.fs.dvc._DVCFileSystem", side_effect=CloneError("SCM error"))
(stage,) = tmp_dir.dvc_gen({"dir": {"foo": "foo", "bar": "bar"}})
imp_stage = dvc.imp(
"https://user:[email protected]/iterative/dvc.git",
"dir",
out="new_dir",
rev="HEAD",
no_exec=True,
)
dvc.push()
shutil.rmtree("dir")
dvc.cache.local.clear()
assert dvc.pull(stage.addressing) == {
"added": [join("dir", "")],
"deleted": [],
"modified": [],
"fetched": 3,
}
with pytest.raises(CloneError, match="SCM error"):
dvc.pull()
with pytest.raises(CloneError, match="SCM error"):
dvc.pull(imp_stage.addressing)
+60
View File
@@ -412,3 +412,63 @@ def test_collect(tmp_dir, dvc):
with pytest.raises(DatasetNotFoundError, match=r"^dataset not found$"):
dvc.datasets["not-existing"]
def test_parametrized(tmp_dir, dvc):
(tmp_dir / "dvc.yaml").dump(
{
"datasets": [
{"name": "${ds1.name}", "url": "${ds1.url}", "type": "dvcx"},
{
"name": "${ds2.name}",
"url": "${ds2.url}",
"type": "dvc",
"path": "${ds2.path}",
},
{
"name": "${ds3.name}",
"url": "${ds3.url}",
"type": "url",
},
]
}
)
(tmp_dir / "params.yaml").dump(
{
"ds1": {"name": "dogs", "url": "dvcx://dogs"},
"ds2": {
"name": "example-get-started",
"url": "[email protected]:iterative/example-get-started.git",
"path": "path",
},
"ds3": {
"name": "cloud-versioning-demo",
"url": "s3://cloud-versioning-demo",
},
}
)
path = (tmp_dir / "dvc.yaml").fs_path
assert dict(dvc.datasets.items()) == {
"dogs": DVCXDataset(
manifest_path=path,
spec=DatasetSpec(name="dogs", url="dvcx://dogs", type="dvcx"),
),
"example-get-started": DVCDataset(
manifest_path=path,
spec=DVCDatasetSpec(
name="example-get-started",
url="[email protected]:iterative/example-get-started.git",
path="path",
type="dvc",
),
),
"cloud-versioning-demo": URLDataset(
manifest_path=path,
spec=DatasetSpec(
name="cloud-versioning-demo",
url="s3://cloud-versioning-demo",
type="url",
),
),
}
+73
View File
@@ -0,0 +1,73 @@
import os
from unittest.mock import ANY
import pytest
from dvc.fs import download
lfs_prefetch_params = [
pytest.param("abc", "abc", id="plain"),
pytest.param(
"*",
"[*]",
marks=pytest.mark.skipif(
os.name == "nt",
reason="forbidden character `*` on Windows filesystem",
),
id="escape-*",
),
pytest.param(
"**",
"[*][*]",
marks=pytest.mark.skipif(
os.name == "nt", reason="forbidden character `*` on Windows filesystem"
),
id="escape-**",
),
pytest.param(
"?",
"[?]",
marks=pytest.mark.skipif(
os.name == "nt", reason="forbidden character `?` on Windows filesystem"
),
id="escape-?",
),
pytest.param("[abc]", "[[]abc]", id="escape-[seq]"),
pytest.param("[!abc]", "[[]!abc]", id="escape-[!seq]"),
]
@pytest.mark.parametrize("dirname, include_name", lfs_prefetch_params)
def test_lfs_prefetch_directory(tmp_dir, dvc, scm, mocker, dirname, include_name):
mock_fetch = mocker.patch("scmrepo.git.lfs.fetch")
tmp_dir.scm_gen(
{
".gitattributes": "data/**/* filter=lfs diff=lfs merge=lfs -text",
f"data/{dirname}/test.txt": "test data",
},
commit="init lfs",
)
rev = scm.get_rev()
with dvc.switch(rev):
download(dvc.dvcfs, f"data/{dirname}", "data")
mock_fetch.assert_called_once_with(
scm, [rev], include=[f"/data/{include_name}/**"], progress=ANY
)
@pytest.mark.parametrize("basename, include_name", lfs_prefetch_params)
def test_lfs_prefetch_file(tmp_dir, dvc, scm, mocker, basename, include_name):
mock_fetch = mocker.patch("scmrepo.git.lfs.fetch")
tmp_dir.scm_gen(
{
".gitattributes": "data/**/* filter=lfs diff=lfs merge=lfs -text",
f"data/{basename}.txt": "test data",
},
commit="init lfs",
)
rev = scm.get_rev()
with dvc.switch(rev):
download(dvc.dvcfs, f"data/{basename}.txt", "data")
mock_fetch.assert_called_once_with(
scm, [rev], include=[f"/data/{include_name}.txt"], progress=ANY
)
+1 -1
View File
@@ -297,7 +297,7 @@ def test_gc_rev_num(tmp_dir, scm, dvc):
def test_date(tmp_dir, scm, dvc):
tmp_dir.dvc_gen("testfile", "content", commit="add testfile")
now = datetime.datetime.now()
now = datetime.datetime.now(tz=datetime.timezone.utc)
datestamp = (now.date() + datetime.timedelta(days=1)).isoformat()
tmp_dir.dvc_gen("testfile", "modified", commit="modified")
+4 -2
View File
@@ -305,11 +305,12 @@ def test_cache_type_is_properly_overridden(tmp_dir, scm, dvc, erepo_dir):
assert scm.is_ignored("foo_imported")
def test_pull_imported_directory_stage(tmp_dir, dvc, erepo_dir):
@pytest.mark.parametrize("dirpath", ["dir", "dir/"])
def test_pull_imported_directory_stage(tmp_dir, dvc, erepo_dir, dirpath):
with erepo_dir.chdir():
erepo_dir.dvc_gen({"dir": {"foo": "foo content"}}, commit="create dir")
dvc.imp(os.fspath(erepo_dir), "dir", "dir_imported")
stage = dvc.imp(os.fspath(erepo_dir), dirpath, "dir_imported")
remove("dir_imported")
dvc.cache.local.clear()
@@ -317,6 +318,7 @@ def test_pull_imported_directory_stage(tmp_dir, dvc, erepo_dir):
dvc.pull(["dir_imported.dvc"])
assert (tmp_dir / "dir_imported").read_text() == {"foo": "foo content"}
assert stage.deps[0].fs_path == "dir"
def test_pull_wildcard_imported_directory_stage(tmp_dir, dvc, erepo_dir):
+6 -3
View File
@@ -1,3 +1,4 @@
import logging
import os
import pytest
@@ -231,13 +232,15 @@ def test_restore_pull(tmp_dir, dvc, run_copy, mocker, local_remote):
assert (tmp_dir / LOCK_FILE).exists()
def test_push_pull_unsupported(tmp_dir, dvc, mocker, run_copy):
def test_push_pull_unsupported(tmp_dir, dvc, mocker, run_copy, local_remote, caplog):
tmp_dir.gen("foo", "foo")
run_copy("foo", "bar", name="copy-foo-bar")
mocker.patch.object(
dvc.cloud, "get_remote_odb", side_effect=RunCacheNotSupported("foo")
)
with pytest.raises(RunCacheNotSupported):
with caplog.at_level(logging.DEBUG):
dvc.push(run_cache=True)
with pytest.raises(RunCacheNotSupported):
assert "failed to push run cache" in caplog.text
with caplog.at_level(logging.DEBUG):
dvc.pull(run_cache=True)
assert "failed to pull run cache" in caplog.text
+13 -1
View File
@@ -163,7 +163,7 @@ def verify_vega_props(plot_id, json_result, title, x, y, **kwargs):
assert len(data) == 1
data = first(data)
assert dpath.get(data, ["content", "title"]) == title
assert dpath.get(data, ["content", "title", "text"]) == title
try:
# TODO confusion_matrix_plot - need to find better way of asserting
@@ -634,3 +634,15 @@ def test_nested_x_defn_collection(tmp_dir, dvc, scm, capsys):
{"step": 1, "Max_Leaf_Nodes": "50", "rev": "workspace"},
{"step": 2, "Max_Leaf_Nodes": "500", "rev": "workspace"},
]
def test_plots_empty_directory(tmp_dir, dvc, scm, capsys):
(tmp_dir / "empty").mkdir()
(tmp_dir / "dvc.yaml").dump({"plots": [{"empty": {}}]})
scm.add(["dvc.yaml"])
scm.commit("commit dvc files")
html_path, _, split_json_result = call(capsys)
assert split_json_result == {}
assert html_path == ""
@@ -1,8 +1,8 @@
import pytest
from dvc_studio_client import env, post_live_metrics
from funcy import first
from dvc.env import (
DVC_EXP_GIT_REMOTE,
DVC_STUDIO_OFFLINE,
DVC_STUDIO_REPO_URL,
DVC_STUDIO_TOKEN,
@@ -11,13 +11,15 @@ from dvc.env import (
from dvc.repo import Repo
from dvc.testing.scripts import COPY_SCRIPT
from dvc.utils.studio import get_subrepo_relpath
from dvc_studio_client import env, post_live_metrics
@pytest.mark.studio
@pytest.mark.parametrize("tmp", [True, False])
@pytest.mark.parametrize("offline", [True, False])
@pytest.mark.parametrize("dvc_exp_git_remote", [None, "DVC_EXP_GIT_REMOTE"])
def test_post_to_studio(
tmp_dir, dvc, scm, exp_stage, mocker, monkeypatch, tmp, offline
tmp_dir, dvc, scm, exp_stage, mocker, monkeypatch, tmp, offline, dvc_exp_git_remote
):
valid_response = mocker.MagicMock()
valid_response.status_code = 200
@@ -28,6 +30,8 @@ def test_post_to_studio(
monkeypatch.setenv(DVC_STUDIO_TOKEN, "STUDIO_TOKEN")
monkeypatch.setenv(DVC_STUDIO_URL, "https://0.0.0.0")
monkeypatch.setenv(DVC_STUDIO_OFFLINE, offline)
if dvc_exp_git_remote:
monkeypatch.setenv(DVC_EXP_GIT_REMOTE, dvc_exp_git_remote)
baseline_sha = scm.get_rev()
exp_rev = first(
@@ -52,7 +56,7 @@ def test_post_to_studio(
assert start_call.kwargs["json"] == {
"type": "start",
"repo_url": "STUDIO_REPO_URL",
"repo_url": dvc_exp_git_remote or "STUDIO_REPO_URL",
"baseline_sha": baseline_sha,
"name": name,
"params": {"params.yaml": {"foo": 1}},
@@ -61,7 +65,7 @@ def test_post_to_studio(
assert done_call.kwargs["json"] == {
"type": "done",
"repo_url": "STUDIO_REPO_URL",
"repo_url": dvc_exp_git_remote or "STUDIO_REPO_URL",
"baseline_sha": baseline_sha,
"name": name,
"client": "dvc",
+1
View File
@@ -1,4 +1,5 @@
import pytest
from dvc_ssh.tests.cloud import SSH, TEST_SSH_KEY_PATH, TEST_SSH_USER
+2
View File
@@ -22,6 +22,7 @@ def test_add(mocker, dvc):
to_remote=False,
remote_jobs=None,
force=False,
relink=True,
)
@@ -45,6 +46,7 @@ def test_add_to_remote(mocker, dvc):
to_remote=True,
remote_jobs=None,
force=False,
relink=True,
)
+4 -4
View File
@@ -1,4 +1,4 @@
from datetime import datetime
from datetime import datetime, timezone
import pytest
@@ -15,7 +15,7 @@ from dvc.cli import main
),
(
{"name": "mydataset", "url": "dvcx://dataset", "type": "dvcx"},
{"rev_lock": "0" * 40, "version": 1, "created_at": datetime.now()},
{"version": 1, "created_at": datetime.now(tz=timezone.utc)},
"Adding mydataset (dvcx://dataset @ v1)\n",
),
(
@@ -65,7 +65,7 @@ def test_add_already_exists(dvc, caplog, mocker):
),
(
{"name": "mydataset", "url": "dvcx://dataset", "type": "dvcx"},
{"rev_lock": "0" * 40, "version": 1, "created_at": datetime.now()},
{"version": 1, "created_at": datetime.now(tz=timezone.utc)},
{"version": 2},
{
"missing": "Updating mydataset (dvcx://dataset @ v2)\n",
@@ -75,7 +75,7 @@ def test_add_already_exists(dvc, caplog, mocker):
),
(
{"name": "mydataset", "url": "dvcx://dataset", "type": "dvcx"},
{"rev_lock": "0" * 40, "version": 2, "created_at": datetime.now()},
{"version": 2, "created_at": datetime.now(tz=timezone.utc)},
{"version": 1},
{
"missing": "Updating mydataset (dvcx://dataset @ v1)\n",
+22 -2
View File
@@ -44,7 +44,21 @@ def test_get(mocker):
def test_get_url(mocker, capsys):
cli_args = parse_args(["get", "repo_url", "src", "--rev", "version", "--show-url"])
cli_args = parse_args(
[
"get",
"repo_url",
"src",
"--rev",
"version",
"--remote",
"myremote",
"--show-url",
"--remote-config",
"k1=v1",
"k2=v2",
]
)
assert cli_args.func == CmdGet
cmd = cli_args.func(cli_args)
@@ -54,4 +68,10 @@ def test_get_url(mocker, capsys):
out, _ = capsys.readouterr()
assert "resource_url" in out
m.assert_called_once_with("src", repo="repo_url", rev="version")
m.assert_called_once_with(
"src",
repo="repo_url",
rev="version",
remote="myremote",
remote_config={"k1": "v1", "k2": "v2"},
)
+2 -2
View File
@@ -166,13 +166,13 @@ def test_experiments_status(dvc, scm, mocker, capsys):
{
"rev": "c61a525a4ff39007301b4516fb6e54b323a0587b",
"name": "I40",
"timestamp": datetime(2022, 6, 9, 20, 49, 48),
"timestamp": datetime(2022, 6, 9, 20, 49, 48), # noqa: DTZ001
"status": "Queued",
},
{
"rev": "8da9c339da30636261a3491a90aafdb760a4168f",
"name": "I60",
"timestamp": datetime(2022, 6, 9, 20, 49, 43),
"timestamp": datetime(2022, 6, 9, 20, 49, 43), # noqa: DTZ001
"status": "Running",
},
]
+10 -1
View File
@@ -1,7 +1,16 @@
from dvc_studio_client.auth import AuthorizationExpiredError
import pytest
from dvc import env
from dvc.cli import main
from dvc.utils.studio import STUDIO_URL
from dvc_studio_client.auth import AuthorizationExpiredError
@pytest.fixture(autouse=True)
def global_config_dir(monkeypatch, tmp_dir_factory):
monkeypatch.setenv(
env.DVC_GLOBAL_CONFIG_DIR, str(tmp_dir_factory.mktemp("studio-login"))
)
def test_studio_login_token_check_failed(mocker):
+3 -3
View File
@@ -1,10 +1,10 @@
import pytest
from dvc_http import HTTPFileSystem, HTTPSFileSystem
from dvc_s3 import S3FileSystem
from dvc_ssh import SSHFileSystem
from dvc.config import RemoteNotFoundError
from dvc.fs import LocalFileSystem, get_cloud_fs, get_fs_cls, get_fs_config
from dvc_http import HTTPFileSystem, HTTPSFileSystem
from dvc_s3 import S3FileSystem
from dvc_ssh import SSHFileSystem
url_cls_pairs = [
("s3://bucket/path", S3FileSystem),
+1 -1
View File
@@ -1,10 +1,10 @@
import pytest
from dvc_s3 import S3FileSystem
from dvc import output
from dvc.fs import LocalFileSystem
from dvc.output import Output
from dvc.stage import Stage
from dvc_s3 import S3FileSystem
@pytest.mark.parametrize(
+2 -2
View File
@@ -1,8 +1,8 @@
import pytest
from dvc_gs import GSFileSystem
from dvc_s3 import S3FileSystem
from dvc.fs import get_cloud_fs
from dvc_gs import GSFileSystem
from dvc_s3 import S3FileSystem
def test_remote_with_hash_jobs(dvc):
+1 -1
View File
@@ -1,7 +1,7 @@
import pytest
from dvc_webdav import WebDAVFileSystem, WebDAVSFileSystem
from dvc.fs import get_cloud_fs
from dvc_webdav import WebDAVFileSystem, WebDAVSFileSystem
from tests.utils.asserts import issubset
url_fmt = "{scheme}://{user}@example.com/public.php/webdav"

Some files were not shown because too many files have changed in this diff Show More