refactor: rename utils.stage to utils.yaml

dvc/dvcfile.py

@@ -17,11 +17,7 @@
 from dvc.stage.loader import SingleStageLoader, StageLoader
 from dvc.utils import relpath
 from dvc.utils.collections import apply_diff
-from dvc.utils.stage import (
-    dump_stage_file,
-    parse_stage,
-    parse_stage_for_update,
-)
+from dvc.utils.yaml import dump_yaml, parse_yaml, parse_yaml_for_update
 
 logger = logging.getLogger(__name__)
 
@@ -101,7 +97,7 @@ def _load(self):
 
         with self.repo.tree.open(self.path) as fd:
             stage_text = fd.read()
-        d = parse_stage(stage_text, self.path)
+        d = parse_yaml(stage_text, self.path)
         self.validate(d, self.relpath)
         return d, stage_text
 
@@ -149,7 +145,7 @@ def dump(self, stage, **kwargs):
         logger.debug(
             "Saving information to '{file}'.".format(file=relpath(self.path))
         )
-        dump_stage_file(self.path, serialize.to_single_stage_file(stage))
+        dump_yaml(self.path, serialize.to_single_stage_file(stage))
         self.repo.scm.track_file(self.relpath)
 
     def remove_with_prompt(self, force=False):
@@ -198,7 +194,7 @@ def _dump_pipeline_file(self, stage):
         data = {}
         if self.exists():
             with open(self.path) as fd:
-                data = parse_stage_for_update(fd.read(), self.path)
+                data = parse_yaml_for_update(fd.read(), self.path)
         else:
             logger.info("Creating '%s'", self.relpath)
             open(self.path, "w+").close()
@@ -214,7 +210,7 @@ def _dump_pipeline_file(self, stage):
         logger.info(
             "Adding stage '%s' to '%s'", stage.name, self.relpath,
         )
-        dump_stage_file(self.path, data)
+        dump_yaml(self.path, data)
         self.repo.scm.track_file(self.relpath)
 
     @property
@@ -243,15 +239,15 @@ def remove_stage(self, stage):
             return
 
         with open(self.path, "r") as f:
-            d = parse_stage_for_update(f.read(), self.path)
+            d = parse_yaml_for_update(f.read(), self.path)
 
         self.validate(d, self.path)
         if stage.name not in d.get("stages", {}):
             return
 
         logger.debug("Removing '%s' from '%s'", stage.name, self.path)
         del d["stages"][stage.name]
-        dump_stage_file(self.path, d)
+        dump_yaml(self.path, d)
 
 
 class Lockfile(FileMixin):
@@ -261,7 +257,7 @@ def load(self):
         if not self.exists():
             return {}
         with self.repo.tree.open(self.path) as fd:
-            data = parse_stage(fd.read(), self.path)
+            data = parse_yaml(fd.read(), self.path)
         try:
             self.validate(data, fname=self.relpath)
         except StageFileFormatError:
@@ -279,14 +275,14 @@ def dump(self, stage, **kwargs):
             open(self.path, "w+").close()
         else:
             with self.repo.tree.open(self.path, "r") as fd:
-                data = parse_stage_for_update(fd.read(), self.path)
+                data = parse_yaml_for_update(fd.read(), self.path)
             modified = data.get(stage.name, {}) != stage_data.get(
                 stage.name, {}
             )
             if modified:
                 logger.info("Updating lock file '%s'", self.relpath)
             data.update(stage_data)
-        dump_stage_file(self.path, data)
+        dump_yaml(self.path, data)
         if modified:
             self.repo.scm.track_file(self.relpath)
 
@@ -295,7 +291,7 @@ def remove_stage(self, stage):
             return
 
         with open(self.path) as f:
-            d = parse_stage_for_update(f.read(), self.path)
+            d = parse_yaml_for_update(f.read(), self.path)
         self.validate(d, self.path)
 
         if stage.name not in d:
@@ -304,7 +300,7 @@ def remove_stage(self, stage):
         logger.debug("Removing '%s' from '%s'", stage.name, self.path)
         del d[stage.name]
 
-        dump_stage_file(self.path, d)
+        dump_yaml(self.path, d)
 
 
 class Dvcfile:

dvc/exceptions.py

@@ -178,12 +178,11 @@ def __init__(self):
         )
 
 
-class StageFileCorruptedError(DvcException):
+class YAMLFileCorruptedError(DvcException):
     def __init__(self, path):
         path = relpath(path)
         super().__init__(
-            "unable to read DVC-file: {} "
-            "YAML file structure is corrupted".format(path)
+            f"unable to read: '{path}', YAML file structure is corrupted"
         )
 
 

dvc/serialize.py

@@ -10,7 +10,7 @@
 from dvc.stage.params import StageParams
 from dvc.stage.utils import resolve_wdir
 from dvc.utils.collections import apply_diff
-from dvc.utils.stage import parse_stage_for_update
+from dvc.utils.yaml import parse_yaml_for_update
 
 if TYPE_CHECKING:
     from dvc.stage import PipelineStage, Stage
@@ -165,7 +165,7 @@ def to_single_stage_file(stage: "Stage"):
     # - apply changes to a returned structure
     # - serialize it
     if stage._stage_text is not None:
-        saved_state = parse_stage_for_update(stage._stage_text, stage.path)
+        saved_state = parse_yaml_for_update(stage._stage_text, stage.path)
         # Stage doesn't work with meta in any way, so .dumpd() doesn't
         # have it. We simply copy it over.
         if "meta" in saved_state:

dvc/stage/cache.py

@@ -11,7 +11,7 @@
 from dvc.stage.loader import StageLoader
 from dvc.utils import dict_sha256, relpath
 from dvc.utils.fs import makedirs
-from dvc.utils.stage import dump_stage_file
+from dvc.utils.yaml import dump_yaml
 
 logger = logging.getLogger(__name__)
 
@@ -154,7 +154,7 @@ def save(self, stage):
         path = self._get_cache_path(cache_key, cache_value)
         dpath = os.path.dirname(path)
         makedirs(dpath, exist_ok=True)
-        dump_stage_file(path, cache)
+        dump_yaml(path, cache)
 
     def is_cached(self, stage):
         return bool(self._load(stage))

dvc/utils/stage.py → dvc/utils/yaml.py

@@ -4,43 +4,43 @@
 from ruamel.yaml import YAML
 from ruamel.yaml.error import YAMLError
 
-from dvc.exceptions import StageFileCorruptedError
+from dvc.exceptions import YAMLFileCorruptedError
 
 try:
     from yaml import CSafeLoader as SafeLoader
 except ImportError:
     from yaml import SafeLoader
 
 
-def load_stage_file(path):
+def load_yaml(path):
     with open(path, encoding="utf-8") as fd:
-        return parse_stage(fd.read(), path)
+        return parse_yaml(fd.read(), path)
 
 
-def parse_stage(text, path):
+def parse_yaml(text, path):
     try:
         return yaml.load(text, Loader=SafeLoader) or {}
     except yaml.error.YAMLError as exc:
-        raise StageFileCorruptedError(path) from exc
+        raise YAMLFileCorruptedError(path) from exc
 
 
-def parse_stage_for_update(text, path):
+def parse_yaml_for_update(text, path):
     """Parses text into Python structure.
 
-    Unlike `parse_stage()` this returns ordered dicts, values have special
+    Unlike `parse_yaml()` this returns ordered dicts, values have special
     attributes to store comments and line breaks. This allows us to preserve
     all of those upon dump.
 
-    This one is, however, several times slower than simple `parse_stage()`.
+    This one is, however, several times slower than simple `parse_yaml()`.
     """
     try:
         yaml = YAML()
         return yaml.load(text) or {}
     except YAMLError as exc:
-        raise StageFileCorruptedError(path) from exc
+        raise YAMLFileCorruptedError(path) from exc
 
 
-def dump_stage_file(path, data):
+def dump_yaml(path, data):
     with open(path, "w", encoding="utf-8") as fd:
         yaml = YAML()
         yaml.default_flow_style = False

tests/func/test_add.py

@@ -17,7 +17,7 @@
     OutputDuplicationError,
     OverlappingOutputPathsError,
     RecursiveAddingWhileUsingFilename,
-    StageFileCorruptedError,
+    YAMLFileCorruptedError,
 )
 from dvc.main import main
 from dvc.output.base import OutputAlreadyTrackedError, OutputIsStageFileError
@@ -27,7 +27,7 @@
 from dvc.system import System
 from dvc.utils import LARGE_DIR_SIZE, file_md5, relpath
 from dvc.utils.fs import path_isin
-from dvc.utils.stage import load_stage_file
+from dvc.utils.yaml import load_yaml
 from tests.basic_env import TestDvc
 from tests.utils import get_gitignore_content
 
@@ -46,7 +46,7 @@ def test_add(tmp_dir, dvc):
     assert stage.outs[0].info["md5"] == md5
     assert stage.md5 is None
 
-    assert load_stage_file("foo.dvc") == {
+    assert load_yaml("foo.dvc") == {
         "outs": [{"md5": "acbd18db4cc2f85cedef654fccc4a4d8", "path": "foo"}],
     }
 
@@ -220,14 +220,14 @@ def test(self):
         ret = main(["add", foo])
         self.assertEqual(ret, 0)
 
-        d = load_stage_file("foo.dvc")
+        d = load_yaml("foo.dvc")
         self.assertEqual(d["outs"][0]["path"], foo)
 
         bar = os.path.join(cwd, self.BAR)
         ret = main(["add", bar])
         self.assertEqual(ret, 0)
 
-        d = load_stage_file("bar.dvc")
+        d = load_yaml("bar.dvc")
         self.assertEqual(d["outs"][0]["path"], self.BAR)
 
 
@@ -371,7 +371,7 @@ def _test(self):
         stage_file = self.data_file_name + DVC_FILE_SUFFIX
         self.assertTrue(os.path.exists(stage_file))
 
-        d = load_stage_file(stage_file)
+        d = load_yaml(stage_file)
         relative_data_path = posixpath.join(
             self.link_name, self.data_file_name
         )
@@ -436,8 +436,7 @@ def test(self):
         assert 1 == ret
 
         expected_error = (
-            "unable to read DVC-file: {} "
-            "YAML file structure is corrupted".format(foo_stage)
+            f"unable to read: '{foo_stage}', YAML file structure is corrupted"
         )
 
         assert expected_error in self._caplog.text
@@ -478,7 +477,7 @@ def test_failed_add_cleanup(tmp_dir, scm, dvc):
     dvc.add("foo")
     tmp_dir.gen("foo.dvc", "- broken\nyaml")
 
-    with pytest.raises(StageFileCorruptedError):
+    with pytest.raises(YAMLFileCorruptedError):
         dvc.add("bar")
 
     assert not os.path.exists("bar.dvc")

tests/func/test_checkout.py

@@ -24,7 +24,7 @@
 from dvc.system import System
 from dvc.utils import relpath
 from dvc.utils.fs import walk_files
-from dvc.utils.stage import dump_stage_file, load_stage_file
+from dvc.utils.yaml import dump_yaml, load_yaml
 from tests.basic_env import TestDvc, TestDvcGit
 from tests.func.test_repro import TestRepro
 from tests.remotes import S3
@@ -221,7 +221,7 @@ def test(self):
         self.assertEqual(0, ret)
 
         stage_path = self.DATA_DIR + DVC_FILE_SUFFIX
-        stage = load_stage_file(stage_path)
+        stage = load_yaml(stage_path)
         staged_files = self.outs_info(stage)
 
         # move instead of remove, to lock inode assigned to stage_files[0].path
@@ -304,10 +304,10 @@ def test(self):
 
 class TestCheckoutMissingMd5InStageFile(TestRepro):
     def test(self):
-        d = load_stage_file(self.file1_stage)
+        d = load_yaml(self.file1_stage)
         del d[Stage.PARAM_OUTS][0][LocalRemote.PARAM_CHECKSUM]
         del d[Stage.PARAM_DEPS][0][LocalRemote.PARAM_CHECKSUM]
-        dump_stage_file(self.file1_stage, d)
+        dump_yaml(self.file1_stage, d)
 
         with pytest.raises(CheckoutError):
             self.dvc.checkout(force=True)

tests/func/test_commit.py

@@ -2,7 +2,7 @@
 
 from dvc.dvcfile import PIPELINE_FILE
 from dvc.stage.exceptions import StageCommitError
-from dvc.utils.stage import dump_stage_file, load_stage_file
+from dvc.utils.yaml import dump_yaml, load_yaml
 
 
 def test_commit_recursive(tmp_dir, dvc):
@@ -63,15 +63,15 @@ def test_commit_changed_md5(tmp_dir, dvc):
     tmp_dir.gen({"file": "file content"})
     (stage,) = dvc.add("file", no_commit=True)
 
-    stage_file_content = load_stage_file(stage.path)
+    stage_file_content = load_yaml(stage.path)
     stage_file_content["md5"] = "1111111111"
-    dump_stage_file(stage.path, stage_file_content)
+    dump_yaml(stage.path, stage_file_content)
 
     with pytest.raises(StageCommitError):
         dvc.commit(stage.path)
 
     dvc.commit(stage.path, force=True)
-    assert "md5" not in load_stage_file(stage.path)
+    assert "md5" not in load_yaml(stage.path)
 
 
 def test_commit_no_exec(tmp_dir, dvc):

tests/func/test_data_cloud.py

@@ -27,7 +27,7 @@
 from dvc.stage.exceptions import StageNotFound
 from dvc.utils import file_md5
 from dvc.utils.fs import remove
-from dvc.utils.stage import dump_stage_file, load_stage_file
+from dvc.utils.yaml import dump_yaml, load_yaml
 from tests.basic_env import TestDvc
 from tests.remotes import (
     GCP,
@@ -523,9 +523,9 @@ def _test(self):
         self.main(["push"])
 
         stage_file_path = stage.relpath
-        content = load_stage_file(stage_file_path)
+        content = load_yaml(stage_file_path)
         del content["outs"][0]["md5"]
-        dump_stage_file(stage_file_path, content)
+        dump_yaml(stage_file_path, content)
 
         with self._caplog.at_level(logging.WARNING, logger="dvc"):
             self._caplog.clear()

tests/func/test_dvcfile.py

@@ -8,7 +8,7 @@
     StageFileFormatError,
 )
 from dvc.stage.loader import StageNotFound
-from dvc.utils.stage import dump_stage_file
+from dvc.utils.yaml import dump_yaml
 
 
 def test_run_load_one_for_multistage(tmp_dir, dvc):
@@ -244,14 +244,14 @@ def test_remove_stage_on_lockfile_format_error(tmp_dir, dvc, run_copy):
     lock_data = lock_file.load()
     lock_data["gibberish"] = True
     data["gibberish"] = True
-    dump_stage_file(lock_file.relpath, lock_data)
+    dump_yaml(lock_file.relpath, lock_data)
     with pytest.raises(StageFileFormatError):
         dvc_file.remove_stage(stage)
 
     lock_file.remove()
     dvc_file.dump(stage)
 
-    dump_stage_file(dvc_file.relpath, data)
+    dump_yaml(dvc_file.relpath, data)
     with pytest.raises(StageFileFormatError):
         dvc_file.remove_stage(stage)