Fix author order comparison and test cleanup

Author: mathpluscode

.claude-plugin/marketplace.json

@@ -13,7 +13,7 @@
       "name": "bibkit",
       "source": "./",
       "description": "A bibliography toolkit for LaTeX",
-      "version": "1.2.1",
+      "version": "1.2.2",
       "keywords": ["bibtex", "bibliography", "latex", "overleaf", "academic", "reference", "citation"],
       "category": "academic",
       "license": "MIT"

.claude-plugin/plugin.json

@@ -1,5 +1,5 @@
 {
   "name": "bibkit",
   "description": "A bibliography toolkit for LaTeX",
-  "version": "1.2.1"
+  "version": "1.2.2"
 }

CLAUDE.md

@@ -7,6 +7,7 @@ A bibliography toolkit for LaTeX — Claude Code plugin.
 ```
 bibkit/
 ├── .claude-plugin/
+│   ├── marketplace.json        ← marketplace catalog
 │   └── plugin.json             ← plugin manifest
 ├── skills/
 │   └── bibtidy/
@@ -36,6 +37,8 @@ bibkit/
 
 ## How it works
 
+### bibtidy
+
 Invoke with `/bibtidy refs.bib`. Claude reads the .bib file, dispatches parallel subagents to verify entries against Google Scholar (WebSearch) and CrossRef (bundled `crossref.py`), then applies fixes sequentially using targeted Edit tool replacements. Every change includes the original entry commented out and a source URL for verification.
 
 ## Versioning

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "bibkit"
-version = "1.2.1"
+version = "1.2.2"
 description = "A bibliography toolkit for LaTeX — Claude Code plugin"
 requires-python = ">=3.10"
 license = "MIT"

skills/bibtidy/SKILL.md

@@ -1,6 +1,8 @@
 ---
 name: bibtidy
 description: Use when the user wants to validate, check, or fix a BibTeX (.bib) reference file — wrong authors, stale arXiv preprints, incorrect metadata, duplicate entries, formatting issues
+argument-hint: <path-to-file.bib>
+allowed-tools: Bash(python3 *), Read, Edit, Agent, WebSearch
 ---
 
 Validate and fix the BibTeX file at: $ARGUMENTS
@@ -28,8 +30,7 @@ All bundled tools live in the `tools/` directory next to this SKILL.md, installe
 ```
 TOOLS_DIR="$HOME/.claude/skills/bibtidy/tools"
 if [ ! -f "$TOOLS_DIR/crossref.py" ]; then
-  echo "Error: bibtidy tools not found. Reinstall the plugin." >&2
-  exit 1
+  TOOLS_DIR="${CLAUDE_PLUGIN_ROOT}/skills/bibtidy/tools"
 fi
 ```
 

skills/bibtidy/tools/compare.py

@@ -44,7 +44,10 @@ def _normalize_title(title: str) -> str:
 
 
 def _normalize_author_list(authors_str: str) -> list[str]:
-    """Parse 'Last, First and Last, First' into sorted lowercase last names."""
+    """Parse 'Last, First and Last, First' into ordered lowercase last names.
+
+    Preserves order so first-author swaps are detected.
+    """
     names = []
     for name in authors_str.split(" and "):
         name = name.strip()
@@ -58,11 +61,11 @@ def _normalize_author_list(authors_str: str) -> list[str]:
         last = last.replace("{", "").replace("}", "")
         last = re.sub(r"\\.", "", last)
         names.append(last.lower())
-    return sorted(names)
+    return names
 
 
 def _crossref_author_last_names(authors: list[str]) -> list[str]:
-    """Extract sorted lowercase last names from CrossRef 'Last, First' strings."""
+    """Extract ordered lowercase last names from CrossRef 'Last, First' strings."""
     names = []
     for a in authors:
         if "," in a:
@@ -71,7 +74,7 @@ def _crossref_author_last_names(authors: list[str]) -> list[str]:
             parts = a.split()
             last = parts[-1].lower() if parts else ""
         names.append(last)
-    return sorted(names)
+    return names
 
 
 def compare_entry(entry: dict, crossref: dict) -> list[dict]:
@@ -92,19 +95,21 @@ def _add(field, bib_val, cr_val, severity="fix"):
         if _normalize_title(bib_title) != _normalize_title(cr_title):
             _add("title", bib_title, cr_title)
 
-    # Authors (compare last names only — first name formats vary)
+    # Authors (compare ordered last names — first name formats vary too much)
     bib_author = entry.get("author", "")
     cr_authors = crossref.get("authors", [])
     if bib_author and cr_authors:
         bib_names = _normalize_author_list(bib_author)
         cr_names = _crossref_author_last_names(cr_authors)
-        # Ignore "others" truncation: only compare if bib has fewer unique names
-        # that don't appear in CrossRef
-        bib_set = set(bib_names)
-        cr_set = set(cr_names)
-        extra_in_bib = bib_set - cr_set
-        missing_from_bib = cr_set - bib_set
-        if extra_in_bib or (missing_from_bib and len(bib_names) >= len(cr_names)):
+        # Compare the overlapping prefix — flag if it disagrees or if
+        # either side has authors the other doesn't.
+        n = min(len(bib_names), len(cr_names))
+        prefix_matches = bib_names[:n] == cr_names[:n]
+        if not prefix_matches:
+            # The authors they both list don't agree (wrong names or order)
+            _add("author", bib_author, " and ".join(cr_authors), "review")
+        elif len(bib_names) != len(cr_names):
+            # One side has more authors than the other — let Claude decide
             _add("author", bib_author, " and ".join(cr_authors), "review")
 
     # Year

tests/conftest.py

@@ -3,12 +3,9 @@
 import os
 import sys
 
-# Add tools directory so tests can import crossref, duplicates, fmt directly
-_tools_dir = os.path.join(os.path.dirname(__file__), "..", "skills", "bibtidy", "tools")
-if _tools_dir not in sys.path:
-    sys.path.insert(0, os.path.abspath(_tools_dir))
+_base = os.path.join(os.path.dirname(__file__), "..")
 
-# Add tests directory so validate.py is importable
-_tests_dir = os.path.dirname(__file__)
-if _tests_dir not in sys.path:
-    sys.path.insert(0, os.path.abspath(_tests_dir))
+for subdir in ("skills/bibtidy/tools", "tests"):
+    d = os.path.abspath(os.path.join(_base, subdir))
+    if d not in sys.path:
+        sys.path.insert(0, d)

tests/test_compare.py

@@ -1,12 +1,6 @@
 #!/usr/bin/env python3
 """Tests for compare.py — field-level comparison between BibTeX and CrossRef."""
 
-import os
-import sys
-
-
-sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", "skills", "bibtidy", "tools"))
-
 from compare import compare_entry
 
 
@@ -57,7 +51,8 @@ def test_different_title(self):
 
 
 class TestAuthors:
-    def test_extra_author_in_bib(self):
+    def test_bib_has_more_authors_than_crossref_is_ok(self):
+        """Bib with more authors is OK — CrossRef may truncate."""
         entry = {"key": "X", "author": "Smith, John and Dayan, Peter"}
         cr = {"authors": ["Smith, John"]}
         ms = compare_entry(entry, cr)
@@ -69,12 +64,47 @@ def test_matching_authors(self):
         ms = compare_entry(entry, cr)
         assert not any(m["field"] == "author" for m in ms)
 
+    def test_swapped_author_order(self):
+        """First-author swap should be flagged."""
+        entry = {"key": "X", "author": "Doe, Jane and Smith, John"}
+        cr = {"authors": ["Smith, John", "Doe, Jane"]}
+        ms = compare_entry(entry, cr)
+        assert any(m["field"] == "author" for m in ms)
+
     def test_bib_has_others(self):
-        """'and others' truncation should not flag missing authors."""
+        """'and others' with full list available should be flagged for Claude to decide."""
         entry = {"key": "X", "author": "Smith, John and others"}
         cr = {"authors": ["Smith, John", "Doe, Jane", "Lee, Bob"]}
         ms = compare_entry(entry, cr)
-        assert not any(m["field"] == "author" for m in ms)
+        assert any(m["field"] == "author" for m in ms)
+
+    def test_bib_has_others_wrong_order(self):
+        """'and others' with wrong prefix order should be flagged."""
+        entry = {"key": "X", "author": "Doe, Jane and Smith, John and others"}
+        cr = {"authors": ["Smith, John", "Doe, Jane", "Lee, Bob"]}
+        ms = compare_entry(entry, cr)
+        assert any(m["field"] == "author" for m in ms)
+
+    def test_bib_has_more_authors_than_crossref(self):
+        """Bib with more authors than CrossRef should be flagged for Claude to decide."""
+        entry = {"key": "X", "author": "Smith, John and Doe, Jane and Lee, Bob"}
+        cr = {"authors": ["Smith, John", "Doe, Jane"]}
+        ms = compare_entry(entry, cr)
+        assert any(m["field"] == "author" for m in ms)
+
+    def test_bib_has_more_but_wrong_prefix(self):
+        """Bib with more authors should be flagged if the shared prefix disagrees."""
+        entry = {"key": "X", "author": "Doe, Jane and Smith, John and Lee, Bob"}
+        cr = {"authors": ["Smith, John", "Doe, Jane"]}
+        ms = compare_entry(entry, cr)
+        assert any(m["field"] == "author" for m in ms)
+
+    def test_bib_missing_authors_no_others(self):
+        """Bib with fewer authors and no 'others' should be flagged."""
+        entry = {"key": "X", "author": "Smith, John"}
+        cr = {"authors": ["Smith, John", "Doe, Jane", "Lee, Bob"]}
+        ms = compare_entry(entry, cr)
+        assert any(m["field"] == "author" for m in ms)
 
 
 class TestYear:

tests/test_crossref.py

@@ -2,14 +2,8 @@
 """Tests for crossref.py — JSON parsing/formatting and error handling."""
 
 import json
-import os
-import sys
-from unittest.mock import patch
-
 import urllib.error
-
-# Ensure the crossref module is importable.
-sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), os.pardir, "skills", "bibtidy", "tools")))
+from unittest.mock import patch
 
 import crossref
 

tests/test_duplicates.py

@@ -6,9 +6,6 @@
 import subprocess
 import sys
 
-
-sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", "skills", "bibtidy", "tools"))
-
 from duplicates import find_duplicates, is_preprint, normalize_title, parse_bib_entries
 
 TOOL_PATH = os.path.join(os.path.dirname(__file__), "..", "skills", "bibtidy", "tools", "duplicates.py")