refactor(metadata): use Metadata.from_email() directly

Replace parse_email() + Metadata.from_raw() with Metadata.from_email()
for simpler code and proper validation. Use parse_metadata() helper
where appropriate and fix empty metadata error handling.

Fixes: #561

Signed-off-by: Lalatendu Mohanty <lmohanty@redhat.com>

Author: LalatenduMohanty

src/fromager/bootstrapper.py

@@ -12,7 +12,6 @@
 import zipfile
 from urllib.parse import urlparse
 
-from packaging.metadata import Metadata, parse_email
 from packaging.requirements import Requirement
 from packaging.utils import NormalizedName, canonicalize_name
 from packaging.version import Version
@@ -907,9 +906,7 @@ def _get_version_from_package_metadata(
             config_settings=pbi.config_settings,
         )
         metadata_filename = source_dir.parent / metadata_dir_base / "METADATA"
-        with open(metadata_filename, "rb") as f:
-            raw_metadata, _ = parse_email(f.read())
-            metadata = Metadata.from_raw(raw_metadata)
+        metadata = dependencies.parse_metadata(metadata_filename)
         return metadata.version
 
     def _resolve_prebuilt_with_history(

src/fromager/candidate.py

@@ -5,7 +5,7 @@
 from io import BytesIO
 from zipfile import ZipFile
 
-from packaging.metadata import Metadata, parse_email
+from packaging.metadata import Metadata
 from packaging.requirements import Requirement
 from packaging.utils import BuildTag, canonicalize_name
 from packaging.version import Version
@@ -14,6 +14,7 @@
 
 logger = logging.getLogger(__name__)
 
+
 @dataclasses.dataclass(frozen=True, order=True, slots=True, repr=False, kw_only=True)
 class Candidate:
     name: str
@@ -109,8 +110,7 @@ def get_metadata_for_wheel(url: str, metadata_url: str | None = None) -> Metadat
             response.raise_for_status()
 
             # Parse metadata directly from the response content
-            raw_metadata, _ = parse_email(response.content)
-            metadata = Metadata.from_raw(raw_metadata)
+            metadata = Metadata.from_email(response.content)
             logger.debug(f"Successfully retrieved metadata via PEP 658 for {url}")
             return metadata
 
@@ -127,10 +127,7 @@ def get_metadata_for_wheel(url: str, metadata_url: str | None = None) -> Metadat
         for n in z.namelist():
             if n.endswith(".dist-info/METADATA"):
                 metadata_content = z.read(n)
-                raw_metadata, _ = parse_email(metadata_content)
-                metadata = Metadata.from_raw(raw_metadata)
-                return metadata
+                return Metadata.from_email(metadata_content)
 
-    # If we didn't find the metadata, return an empty metadata object
-    raw_metadata, _ = parse_email(b"")
-    return Metadata.from_raw(raw_metadata)
+    # If we didn't find the metadata, raise an error
+    raise ValueError(f"Could not find METADATA file in wheel: {url}")

tests/test_pep658_support.py

@@ -56,10 +56,11 @@ def test_get_metadata_with_pep658_success(self, mock_session) -> None:
         metadata = get_metadata_for_wheel(wheel_url, metadata_url)
 
         # Verify the metadata was parsed correctly
-        assert metadata["Name"] == "test-package"
-        assert metadata["Version"] == "1.0.0"
-        assert metadata["Summary"] == "A test package"
-        assert "requests >= 2.0.0" in metadata.get_all("Requires-Dist", [])
+        assert metadata.name == "test-package"
+        assert str(metadata.version) == "1.0.0"
+        assert metadata.summary == "A test package"
+        assert metadata.requires_dist is not None
+        assert any(str(req) == "requests>=2.0.0" for req in metadata.requires_dist)
 
         # Verify only the metadata URL was called, not the wheel URL
         mock_session.get.assert_called_once_with(metadata_url)