Support parsing metadata filenames for purls

Supports parsing .dsc, copyright and changelog files, typically present in the debian package/metadata archives for name and version. Signed-off-by: Ayan Sinha Mahapatra <[email protected]>
aboutcode-org · Jan 30, 2024 · 743841c · 743841c
1 parent 240e24a
commit 743841c
Show file tree

Hide file tree

Showing 2 changed files with 41 additions and 2 deletions.
diff --git a/src/debian_inspector/package.py b/src/debian_inspector/package.py
@@ -93,7 +93,8 @@ def from_filename(cls, filename):
         return cls(
             name=name,
             version=version,
-            original_filename=filename)
+            original_filename=filename
+        )
 
     def to_dict(self):
         data = {}
@@ -110,16 +111,36 @@ def to_tuple(self):
         return tuple(v for v in self.to_dict().values() if v != 'original_filename')
 
 
+@attrs
+class CodeMetadata(CodeArchive):
+    """
+    A .dsc, copyright or changelog file present in the debian
+    package/metadata archive and contains package information
+    on the filename or as file contents.
+
+    For example in ./changelogs/main/d/diffutils/ there are
+    files such as:
+    - diffutils_3.7-5_copyright
+    - diffutils_3.7-5_changelog
+    And in .pool/main/b/base-files/ there are files such as:
+    - base-files_11.1+deb11u8.dsc
+    """
+
+
 def get_nva(filename):
     """
     Return a tuple of (name string, Version object, archictecture string or
     None) parsed from the `filename` of .deb, .udeb, .orig or .debian archive..
     """
     is_known = False
-    if filename.endswith(('.deb', '.udeb')):
+    if filename.endswith(('.deb', '.udeb', '.dsc')):
         basename, _extension = path.splitext(filename)
         is_known = True
 
+    elif filename.endswith(('_changelog', '_copyright')):
+        basename, _, _ = filename.rpartition("_")
+        is_known = True
+
     elif filename.endswith(('.tar.gz', '.tar.xz', '.tar.bz2', '.tar.lzma')):
         # A Format: 3.0 archive.
         # Note that we ignore the legacy .diff.gz files for Format: 1.0

diff --git a/tests/test_package.py b/tests/test_package.py
@@ -121,6 +121,24 @@ def test_CodeArchive_from_filename(self):
             original_filename=fn)
         assert debarch == expected
 
+    def test_CodeMetadata_from_filename_dsc(self):
+        fn = 'base-files_11.1+deb11u8.dsc'
+        debarch = package.CodeMetadata.from_filename(fn)
+        expected = package.CodeMetadata(
+            name='base-files',
+            version=version.Version(epoch=0, upstream='11.1+deb11u8', revision='0'),
+            original_filename=fn)
+        assert debarch == expected
+
+    def test_CodeMetadata_from_filename_copyright(self):
+        fn = 'bash_4.1-3+deb6u2_copyright'
+        debarch = package.CodeMetadata.from_filename(fn)
+        expected = package.CodeMetadata(
+            name='bash',
+            version=version.Version(epoch=0, upstream='4.1', revision='3+deb6u2'),
+            original_filename=fn)
+        assert debarch == expected
+
     def test_CodeArchive_from_filename_supports_tar_gz_bz2_and_xz(self):
         package.CodeArchive.from_filename('python2.7_2.7.3-0ubuntu3.4.orig.tar.gz')
         package.CodeArchive.from_filename('python2.7_2.7.3-0ubuntu3.4.debian.tar.gz')