Stone_Red/dvc
1
0
Fork 0
mirror of https://github.com/Stone-Red-Code/dvc.git synced 2026-11-02 07:18:25 +01:00
Code Issues Packages Projects Releases Wiki Activity

remote: set hash_name as md5

Browse Source 
Fixes #7977
This commit is contained in:
Ruslan Kuprieiev
2023-03-01 13:26:24 +02:00
parent 1a7ae16f69
commit 59a4c4f67c
4 changed files with 83 additions and 79 deletions
Download Patch File Download Diff File Expand all files Collapse all files
dvc/data_cloud.py
+1 -1
View File
@@ -35,7 +35,7 @@ class Remote:
if self.worktree:
path = self.fs.path.join(path, ".dvc", "cache")
return get_odb(self.fs, path, **self.config)
return get_odb(self.fs, path, hash_name="md5", **self.config)
class DataCloud:
dvc/testing/workspace_tests.py
+82
View File
@@ -282,3 +282,85 @@ class TestGetUrl:
def test_get_url_nonexistent(self, cloud):
with pytest.raises(URLMissingError):
Repo.get_url(str(cloud / "nonexistent"), config=cloud.config)
class TestToRemote:
def test_add_to_remote(self, tmp_dir, dvc, remote, workspace):
workspace.gen("foo", "foo")
url = "remote://workspace/foo"
[stage] = dvc.add(url, to_remote=True)
assert not (tmp_dir / "foo").exists()
assert (tmp_dir / "foo.dvc").exists()
assert len(stage.deps) == 0
assert len(stage.outs) == 1
hash_info = stage.outs[0].hash_info
meta = stage.outs[0].meta
assert hash_info.name == "md5"
assert hash_info.value == "acbd18db4cc2f85cedef654fccc4a4d8"
assert (remote / "ac" / "bd18db4cc2f85cedef654fccc4a4d8").read_text() == "foo"
assert meta.size == len("foo")
def test_import_url_to_remote_file(self, tmp_dir, dvc, workspace, remote):
workspace.gen("foo", "foo")
url = "remote://workspace/foo"
stage = dvc.imp_url(url, to_remote=True)
assert stage.deps[0].hash_info.value is not None
assert not (tmp_dir / "foo").exists()
assert (tmp_dir / "foo.dvc").exists()
assert len(stage.deps) == 1
assert stage.deps[0].def_path == url
assert len(stage.outs) == 1
hash_info = stage.outs[0].hash_info
assert hash_info.name == "md5"
assert hash_info.value == "acbd18db4cc2f85cedef654fccc4a4d8"
assert (remote / "ac" / "bd18db4cc2f85cedef654fccc4a4d8").read_text() == "foo"
assert stage.outs[0].meta.size == len("foo")
def test_import_url_to_remote_dir(self, tmp_dir, dvc, workspace, remote):
import json
workspace.gen(
{
"data": {
"foo": "foo",
"bar": "bar",
"sub_dir": {"baz": "sub_dir/baz"},
}
}
)
url = "remote://workspace/data"
stage = dvc.imp_url(url, to_remote=True)
assert not (tmp_dir / "data").exists()
assert (tmp_dir / "data.dvc").exists()
assert len(stage.deps) == 1
assert stage.deps[0].def_path == url
assert len(stage.outs) == 1
hash_info = stage.outs[0].hash_info
assert hash_info.name == "md5"
assert hash_info.value == "55d05978954d1b2cd7b06aedda9b9e43.dir"
file_parts = json.loads(
(remote / "55" / "d05978954d1b2cd7b06aedda9b9e43.dir").read_text()
)
assert len(file_parts) == 3
assert {file_part["relpath"] for file_part in file_parts} == {
"foo",
"bar",
"sub_dir/baz",
}
for file_part in file_parts:
md5 = file_part["md5"]
assert (remote / md5[:2] / md5[2:]).read_text() == file_part["relpath"]
tests/func/test_add.py
-20
View File
@@ -898,26 +898,6 @@ def test_add_long_fname(tmp_dir, dvc):
assert (tmp_dir / "data").read_text() == {name: "foo"}
def test_add_to_remote(tmp_dir, dvc, remote, workspace):
workspace.gen("foo", "foo")
url = "remote://workspace/foo"
[stage] = dvc.add(url, to_remote=True)
assert not (tmp_dir / "foo").exists()
assert (tmp_dir / "foo.dvc").exists()
assert len(stage.deps) == 0
assert len(stage.outs) == 1
hash_info = stage.outs[0].hash_info
meta = stage.outs[0].meta
with open(remote.oid_to_path(hash_info.value), encoding="utf-8") as stream:
assert stream.read() == "foo"
assert meta.size == len("foo")
def test_add_to_remote_absolute(tmp_dir, make_tmp_dir, dvc, remote):
tmp_abs_dir = make_tmp_dir("abs")
tmp_foo = tmp_abs_dir / "foo"
tests/func/test_import_url.py
-58
View File
@@ -1,4 +1,3 @@
import json
import os
import textwrap
from uuid import uuid4
@@ -171,63 +170,6 @@ def test_import_url_preserve_fields(tmp_dir, dvc):
)
def test_import_url_to_remote_single_file(tmp_dir, dvc, workspace, local_remote):
workspace.gen("foo", "foo")
url = "remote://workspace/foo"
stage = dvc.imp_url(url, to_remote=True)
assert stage.deps[0].hash_info.value is not None
assert not (tmp_dir / "foo").exists()
assert (tmp_dir / "foo.dvc").exists()
assert len(stage.deps) == 1
assert stage.deps[0].def_path == url
assert len(stage.outs) == 1
hash_info = stage.outs[0].hash_info
with open(local_remote.oid_to_path(hash_info.value), encoding="utf-8") as fobj:
assert fobj.read() == "foo"
assert stage.outs[0].meta.size == len("foo")
def test_import_url_to_remote_directory(tmp_dir, dvc, workspace, local_remote):
workspace.gen(
{
"data": {
"foo": "foo",
"bar": "bar",
"sub_dir": {"baz": "sub_dir/baz"},
}
}
)
url = "remote://workspace/data"
stage = dvc.imp_url(url, to_remote=True)
assert not (tmp_dir / "data").exists()
assert (tmp_dir / "data.dvc").exists()
assert len(stage.deps) == 1
assert stage.deps[0].def_path == url
assert len(stage.outs) == 1
hash_info = stage.outs[0].hash_info
with open(local_remote.oid_to_path(hash_info.value), encoding="utf-8") as stream:
file_parts = json.load(stream)
assert len(file_parts) == 3
assert {file_part["relpath"] for file_part in file_parts} == {
"foo",
"bar",
"sub_dir/baz",
}
for file_part in file_parts:
with open(local_remote.oid_to_path(file_part["md5"]), encoding="utf-8") as fobj:
assert fobj.read() == file_part["relpath"]
def test_import_url_to_remote_absolute(tmp_dir, make_tmp_dir, dvc, local_remote):
tmp_abs_dir = make_tmp_dir("abs")
tmp_foo = tmp_abs_dir / "foo"