Fix author detection for compact dotted names (#4229)

src/cluecode/copyrights.py

@@ -431,6 +431,22 @@ def get_tokens(numbered_lines, splitter=re.compile(r'[\t =;]+').split):
         # space before in some cases: exclude (digit , (s , (c , (-
         # this allows to recover from words like KISA(Korean
         line = re.sub(pattern=r'(\([^rsc\-\d])', repl=r' \g<1>', string=line)
+
+        # Split compact author prefixes such as Author:Frankie.Chu while
+        # preserving colons elsewhere, for example in URLs.
+        line = re.sub(
+            pattern=r'(^|\s)([Aa]uthor):(?=\S)',
+            repl=r'\1\2 ',
+            string=line,
+        )
+
+        # Normalize dotted names only when they follow an Author tag.
+        line = re.sub(
+            pattern=r'(^|\s)([Aa]uthor)\s+([A-Z][a-z]+)\.([A-Z][a-z]+)(?=\b|[,;])',
+            repl=r'\1\2 \3 \4',
+            string=line,
+        )
+
         for tok in splitter(line):
             # strip trailing quotes+comma
             if tok.endswith("',"):

tests/cluecode/test_copyrights_basic.py

@@ -319,6 +319,22 @@ def test_detect_with_lines_only_holders(self):
         ))
         assert results == expected
 
+    def test_detect_author_without_space_and_with_dotted_name(self):
+        numbered_lines = [
+            (1, '//  Date:9 April,2012'),
+            (2, '//  Author:Frankie.Chu'),
+        ]
+        expected = [
+            copyrights.AuthorDetection('Frankie Chu', 2, 2),
+        ]
+        results = list(copyrights.detect_copyrights_from_lines(
+            numbered_lines,
+            include_copyrights=False,
+            include_holders=False,
+            include_authors=True,
+        ))
+        assert results == expected
+
 
 def check_full_detections(expected, test_file):
     """