Refactor parse_modular, docstrings and small type annotations

- Pass only necesary data to parse_modular (the path string)

[noissue]

pulp_rpm/app/constants.py

@@ -172,7 +172,7 @@
     CONTEXT="module_context",
     EOL="eol_date",
     OBSOLETE_BY_MODULE="obsoleted_by_module_name",
-    OBSOLETE_BY_STREAM="obsoleted_by_module_name",
+    OBSOLETE_BY_STREAM="obsoleted_by_module_stream",
 )
 
 # Mandatory fields for Modulemd types

pulp_rpm/app/modulemd.py

@@ -62,13 +62,16 @@ def modules_packages(modules):
     version.add_content(Package.objects.filter(pk__in=packages_to_add))
 
 
-def split_modulemd_file(file):
+def split_modulemd_file(file: str):
     """
     Helper method to preserve original formatting of modulemd.
+
+    Args:
+        file: Absolute path to file
     """
     with tempfile.TemporaryDirectory(dir=".") as tf:
         decompressed_path = os.path.join(tf, "modulemd.yaml")
-        cr.decompress_file(file.path, decompressed_path, cr.AUTO_DETECT_COMPRESSION)
+        cr.decompress_file(file, decompressed_path, cr.AUTO_DETECT_COMPRESSION)
         with open(decompressed_path) as modulemd_file:
             for doc in modulemd_file.read().split("---"):
                 # strip any spaces or newlines from either side, strip the document end marking,
@@ -94,7 +97,7 @@ def check_mandatory_module_fields(module, required_fields):
 
 def create_modulemd(modulemd, snippet):
     """
-    Create dict with modulemd data to can be saved to DB.
+    Create dict with modulemd data to be saved to DB.
     """
     new_module = dict()
     new_module[PULP_MODULE_ATTR.NAME] = modulemd["data"].get("name")
@@ -176,9 +179,12 @@ def create_modulemd_obsoletes(obsolete, snippet):
     return new_obsolete
 
 
-def parse_modular(file):
+def parse_modular(file: str):
     """
     Parse all modular metadata.
+
+    Args:
+        file: Absolute path to file
     """
     modulemd_all = []
     modulemd_defaults_all = []
@@ -215,7 +221,14 @@ def parse_modular(file):
 
 
 class ModularYamlLoader(yaml.SafeLoader):
-    """Custom Loader that preserve unquoted float in specific fields"""
+    """
+    Custom Loader that preserve unquoted float in specific fields (see #3285).
+
+    Motivation (for customizing YAML parsing) is that libmodulemd also implement safe-quoting:
+    https://github.com/fedora-modularity/libmodulemd/blob/main/modulemd/tests/test-modulemd-quoting.c
+
+    This class is based on https://stackoverflow.com/a/74334992
+    """
 
     # Field to preserve (will bypass yaml casting)
     PRESERVED_FIELDS = ("name", "stream", "version", "context", "arch")

pulp_rpm/app/tasks/synchronizing.py

@@ -970,8 +970,13 @@ async def parse_repository_metadata(self, repomd, metadata_results):
                 await self.put(dc)
 
     async def parse_modules_metadata(self, modulemd_result):
-        """Parse modules' metadata which define what packages are built for specific releases."""
-        modulemd_all, defaults_all, obsoletes_all = parse_modular(modulemd_result)
+        """
+        Parse modules' metadata which define what packages are built for specific releases.
+
+        Args:
+            modulemd_result(pulpcore.download.base.DownloadResult): downloaded modulemd file
+        """
+        modulemd_all, defaults_all, obsoletes_all = parse_modular(modulemd_result.path)
 
         modulemd_dcs = []
 

pulp_rpm/tests/unit/test_modulemd.py

@@ -1,5 +1,4 @@
 from pulp_rpm.app.modulemd import parse_modular
-from types import SimpleNamespace
 import os
 
 sample_file_data = """
@@ -90,10 +89,9 @@ def test_parse_modular_preserves_literal_unquoted_values_3285(tmp_path):
     with open(file_name, "w") as file:
         file.write(sample_file_data)
 
-    mock_obj = SimpleNamespace(path=str(file_name))
-    all, defaults, obsoletes = parse_modular(mock_obj)
+    all, defaults, obsoletes = parse_modular(file_name)
 
-    # check the first two entries
+    # check normal,defaults and obsoletes modulemds
     kangoroo1 = all[0]  # unquoted
     kangoroo2 = all[1]  # quoted
     modulemd_defaults = defaults[0]
@@ -115,10 +113,18 @@ def test_parse_modular_preserves_literal_unquoted_values_3285(tmp_path):
 
     # 'stream' keys which have non-scalar values (e.g. list) are parsed normally.
     # Otherwise, weird results are produced (internal pyyaml objects)
+    assert modulemd_defaults["module"] == "avocado"
+    assert modulemd_defaults.get("modified") is None  # not present
     assert modulemd_defaults["profiles"]["stream"] == ["default"]
     assert modulemd_defaults["profiles"]["5.30"] == ["default"]
 
     # parse_modular changes the structure and key names for obsoletes
+    assert modulemd_obsoletes["modified"] == "2022-01-24T08:54Z"
     assert modulemd_obsoletes["module_name"] == "perl"
     assert modulemd_obsoletes["module_stream"] == "5.30"
-    assert modulemd_obsoletes["obsoleted_by_module_name"] == "5.40"
+    assert modulemd_obsoletes["eol_date"] == "2021-06-01T00:00Z"
+    assert (
+        modulemd_obsoletes["message"]
+        == "Module stream perl:5.30 is no longer supported. Please switch to perl:5.32"
+    )
+    assert modulemd_obsoletes["obsoleted_by_module_stream"] == "5.40"