data cloud: refactor verify checksums test

dvc/data_cloud.py

@@ -89,8 +89,7 @@ class DataCloud(object):
             if self.repo.cache.local.exists(cache_file):
                 # We can safely save here, as existing corrupted files will be
                 # removed upon status, while files corrupted during download
-                # will not be copied from tmp_file (see download
-                # implementation in `remote/base`)
+                # will not be moved from tmp_file (see `RemoteBASE.download()`)
                 self.repo.state.save(cache_file, checksum)
 
     def status(self, cache, jobs=None, remote=None, show_checksums=False):

tests/func/test_data_cloud.py

@@ -9,10 +9,11 @@ import pytest
 from mock import patch
 
 from dvc.cache import NamedCache
+from dvc.compat import fspath
 from dvc.config import Config
 from dvc.data_cloud import DataCloud
 from dvc.main import main
-from dvc.remote import RemoteAZURE
+from dvc.remote import RemoteAZURE, RemoteConfig
 from dvc.remote import RemoteGDrive
 from dvc.remote import RemoteGS
 from dvc.remote import RemoteHDFS
@@ -653,3 +654,37 @@ class TestShouldWarnOnNoChecksumInLocalAndRemoteCache(TestDvc):
         assert self.message_header in self._caplog.text
         assert self.message_foo_part in self._caplog.text
         assert self.message_bar_part in self._caplog.text
+
+
+def test_verify_checksums(tmp_dir, scm, dvc, mocker, tmp_path_factory):
+    tmp_dir.dvc_gen({"file": "file1 content"}, commit="add file")
+    tmp_dir.dvc_gen({"dir": {"subfile": "file2 content"}}, commit="add dir")
+
+    RemoteConfig(dvc.config).add(
+        "local_remote",
+        fspath(tmp_path_factory.mktemp("local_remote")),
+        default=True,
+    )
+    dvc.push()
+
+    # remove artifacts and cache to trigger fetching
+    os.remove("file")
+    shutil.rmtree("dir")
+    shutil.rmtree(dvc.cache.local.cache_dir)
+
+    checksum_spy = mocker.spy(dvc.cache.local, "get_file_checksum")
+
+    dvc.pull()
+    assert checksum_spy.call_count == 0
+
+    # Removing cache will invalidate existing state entries
+    shutil.rmtree(dvc.cache.local.cache_dir)
+
+    dvc.config.set(
+        Config.SECTION_REMOTE_FMT.format("local_remote"),
+        Config.SECTION_REMOTE_VERIFY,
+        "True",
+    )
+
+    dvc.pull()
+    assert checksum_spy.call_count == 3

tests/func/test_remote.py

@@ -13,7 +13,6 @@ from dvc.path_info import PathInfo
 from dvc.remote import RemoteLOCAL, RemoteConfig
 from dvc.remote.base import RemoteBASE
 from dvc.compat import fspath
-from dvc.scm import Git
 from tests.basic_env import TestDvc
 from tests.remotes import Local
 
@@ -258,33 +257,3 @@ def test_modify_missing_remote(dvc):
 
     with pytest.raises(ConfigError, match=r"unable to find remote section"):
         remote_config.modify("myremote", "gdrive_client_id", "xxx")
-
-
-def test_verify_checksums(tmp_dir, erepo_dir, mocker):
-    with erepo_dir.chdir():
-        erepo_dir.dvc_gen({"file": "file1 content"}, commit="add file")
-        erepo_dir.dvc_gen(
-            {"dir": {"subfile": "file2 content"}}, commit="add " "dir"
-        )
-
-    Git.clone(fspath(erepo_dir), fspath(tmp_dir))
-
-    from dvc.repo import Repo
-
-    dvc = Repo(fspath(tmp_dir))
-
-    file_md5_spy = mocker.spy(dvc.cache.local, "get_file_checksum")
-
-    dvc.pull()
-    assert file_md5_spy.call_count == 0
-
-    # Removing cache will invalidate existing state entries
-    shutil.rmtree(dvc.cache.local.cache_dir)
-    dvc.config.set(
-        Config.SECTION_REMOTE_FMT.format("upstream"),
-        Config.SECTION_REMOTE_VERIFY,
-        "True",
-    )
-
-    dvc.pull()
-    assert file_md5_spy.call_count == 3