feat: Implement conference name normalization to reduce 'unknown' assessments (fixes #69) (#85)

This change adds normalization of conference names during BibTeX parsing
to reduce variations caused by different citation styles, particularly
the "Proceedings of" and "Proceedings of the" prefixes that appear
inconsistently across BibTeX entries.

Changes:
- Add _normalize_conference_name() method to BibtexParser
- Apply normalization to all conference name extraction paths
- Remove "Proceedings of the" and "Proceedings of" prefixes (case-insensitive)
- Normalize whitespace in conference names

Benefits:
- Reduces duplicate queries for the same conference with different formatting
- Improves cache hit rate for case-insensitive conference name matching
- Works with existing InputNormalizer infrastructure for additional normalization

Testing:
- Added test_conference_name_normalization() to verify BibTeX parsing
- Added test_normalize_conference_name_method() to verify normalization logic
- All 272 tests pass with 57% coverage

[AI-assisted]

Co-authored-by: florath-ai-assistant[bot] <Andreas.Florath@telekom.de>

Author: coding-ai-assistant[bot]

src/aletheia_probe/bibtex_parser.py

@@ -249,7 +249,7 @@ def _extract_conference_name(entry: Entry) -> str | None:
         """Extract conference name from a BibTeX @inproceedings entry.
 
         Prioritizes series name (normalized) over full booktitle.
-        Attempts to extract conference series by removing years and ordinals.
+        Normalizes conference names to improve matching across variations.
 
         Args:
             entry: BibTeX entry
@@ -263,20 +263,51 @@ def _extract_conference_name(entry: Entry) -> str | None:
             # Remove common artifacts (quotes, extra spaces)
             series = series.strip("'\"").strip()
             if series:
-                return series
+                return BibtexParser._normalize_conference_name(series)
 
         # Priority 2: Try to extract from booktitle
         booktitle = BibtexParser._get_field_safely(entry, "booktitle")
         if booktitle:
-            return booktitle
+            return BibtexParser._normalize_conference_name(booktitle)
 
         # Priority 3: Fallback to organization
         organization = BibtexParser._get_field_safely(entry, "organization")
         if organization:
-            return organization
+            return BibtexParser._normalize_conference_name(organization)
 
         return None
 
+    @staticmethod
+    def _normalize_conference_name(name: str) -> str:
+        """Normalize conference names to improve matching across variations.
+
+        Removes common prefixes like "Proceedings of" and "Proceedings of the"
+        to reduce variation between different citation styles.
+
+        Args:
+            name: Raw conference name
+
+        Returns:
+            Normalized conference name
+
+        Examples:
+            "Proceedings of the IEEE conference on computer vision" ->
+            "IEEE conference on computer vision"
+            "Proceedings of Semantic Web" -> "Semantic Web"
+        """
+        import re
+
+        # Remove "Proceedings of the" prefix (case-insensitive)
+        name = re.sub(r"^proceedings\s+of\s+the\s+", "", name, flags=re.IGNORECASE)
+
+        # Remove "Proceedings of" prefix (case-insensitive)
+        name = re.sub(r"^proceedings\s+of\s+", "", name, flags=re.IGNORECASE)
+
+        # Clean up any extra whitespace
+        name = re.sub(r"\s+", " ", name).strip()
+
+        return name
+
     @staticmethod
     def _extract_authors_safely(entry: Entry) -> str | None:
         """Extract author information from a BibTeX entry with error handling.

tests/unit/test_bibtex_parser.py

@@ -768,7 +768,8 @@ def test_parse_bibtex_conference_with_nested_braces(self, tmp_path):
         )
         assert entry.title == expected_title
 
-        # Conference name should have nested braces removed
+        # Conference name should have nested braces removed and no proceedings prefix
+        # Note: The booktitle doesn't have "Proceedings of" so it stays as is
         expected_conference = (
             "2018 IEEE 11th International Conference on Cloud Computing (CLOUD)"
         )
@@ -916,3 +917,97 @@ def test_parse_bibtex_file_with_latex_escapes(self, tmp_path):
         # Check third entry with multiple escapes
         entry3 = [e for e in entries if e.key == "test_multiple"][0]
         assert entry3.journal_name == 'Test & Review: "Quality" % Assessment'
+
+    def test_conference_name_normalization(self, tmp_path):
+        """Test normalization of conference names to reduce variations."""
+        bibtex_content = """
+@inproceedings{cvpr_proceedings,
+    title={Test Paper 1},
+    booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
+    author={Author One},
+    year={2023}
+}
+
+@inproceedings{cvpr_short,
+    title={Test Paper 2},
+    booktitle={IEEE Conference on Computer Vision and Pattern Recognition},
+    author={Author Two},
+    year={2023}
+}
+
+@inproceedings{neurips_proceedings,
+    title={Test Paper 3},
+    booktitle={Proceedings of Advances in Neural Information Processing Systems},
+    author={Author Three},
+    year={2023}
+}
+
+@inproceedings{neurips_normal,
+    title={Test Paper 4},
+    booktitle={Advances in Neural Information Processing Systems},
+    author={Author Four},
+    year={2023}
+}
+"""
+        test_file = tmp_path / "test_conference_normalization.bib"
+        test_file.write_text(bibtex_content, encoding="utf-8")
+
+        entries = BibtexParser.parse_bibtex_file(test_file)
+
+        assert len(entries) == 4
+
+        # Both CVPR entries should normalize to the same form (without "Proceedings of the")
+        cvpr_proceedings = [e for e in entries if e.key == "cvpr_proceedings"][0]
+        cvpr_short = [e for e in entries if e.key == "cvpr_short"][0]
+        assert (
+            cvpr_proceedings.journal_name
+            == "IEEE Conference on Computer Vision and Pattern Recognition"
+        )
+        assert (
+            cvpr_short.journal_name
+            == "IEEE Conference on Computer Vision and Pattern Recognition"
+        )
+
+        # Both NeurIPS entries should normalize to the same form (without "Proceedings of")
+        neurips_proceedings = [e for e in entries if e.key == "neurips_proceedings"][0]
+        neurips_normal = [e for e in entries if e.key == "neurips_normal"][0]
+        assert (
+            neurips_proceedings.journal_name
+            == "Advances in Neural Information Processing Systems"
+        )
+        assert (
+            neurips_normal.journal_name
+            == "Advances in Neural Information Processing Systems"
+        )
+
+    def test_normalize_conference_name_method(self):
+        """Test the _normalize_conference_name static method directly."""
+        # Test "Proceedings of the" removal
+        result = BibtexParser._normalize_conference_name(
+            "Proceedings of the IEEE Conference on Computer Vision"
+        )
+        assert result == "IEEE Conference on Computer Vision"
+
+        # Test "Proceedings of" removal
+        result = BibtexParser._normalize_conference_name(
+            "Proceedings of Advances in Neural Information Processing Systems"
+        )
+        assert result == "Advances in Neural Information Processing Systems"
+
+        # Test case-insensitive matching
+        result = BibtexParser._normalize_conference_name(
+            "PROCEEDINGS OF THE International Conference on Machine Learning"
+        )
+        assert result == "International Conference on Machine Learning"
+
+        # Test name without "Proceedings of" prefix remains unchanged
+        result = BibtexParser._normalize_conference_name(
+            "IEEE Conference on Computer Vision and Pattern Recognition"
+        )
+        assert result == "IEEE Conference on Computer Vision and Pattern Recognition"
+
+        # Test whitespace normalization
+        result = BibtexParser._normalize_conference_name(
+            "Proceedings  of  the   Conference    Name"
+        )
+        assert result == "Conference Name"