From 4bbb16f348f6ae4cb9541930fec2210335ce524e Mon Sep 17 00:00:00 2001
From: ZHIJUN XU <57060481+xzjncu@users.noreply.github.com>
Date: Mon, 27 Jul 2026 09:23:11 +0800
Subject: [PATCH] fix: normalize_title retains Unicode alphanumeric chars via
 NFKC+casefold; add full CJK name equality check for leading author matching

- normalize_title(): replace ASCII-only a-z0-9 filter with unicodedata.normalize("NFKC") + character.isalnum() to retain CJK/Unicode characters
- _full_leading_author_matches(): add NFKC+casefold full-name equality check before falling back to per-part comparison
- _leading_author_status(): add normalized full-name equality as a match condition alongside family-name key comparison

Fixes source_pdf_mismatch false negatives for Chinese-titled papers (e.g. same title/author previously scored title_similarity=0.0, leading_author=conflict).
---
 skills/deeppapernote/scripts/common.py | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/skills/deeppapernote/scripts/common.py b/skills/deeppapernote/scripts/common.py
index e3ecd10..0e73b22 100644
--- a/skills/deeppapernote/scripts/common.py
+++ b/skills/deeppapernote/scripts/common.py
@@ -135,7 +135,10 @@ def strip_tags(text: str) -> str:
 
 
 def normalize_title(text: str) -> str:
-    return re.sub(r"[^a-z0-9\s]", "", normalize_whitespace(text).lower()).strip()
+    normalized = unicodedata.normalize("NFKC", normalize_whitespace(text)).casefold()
+    return normalize_whitespace("".join(
+        character for character in normalized if character.isalnum() or character.isspace()
+    ))
 
 
 LOCAL_PDF_PREFIX_PATTERN = re.compile(r"^(?:[^-]{1,120})\s+-\s+(?:19|20)\d{2}\s+-\s+")
@@ -576,6 +579,10 @@ def _full_leading_author_matches(
     work_authors = _dedupe_string_list(work_record.get("authors", []))
     if not source_authors or not work_authors:
         return False
+    source_normalized = unicodedata.normalize("NFKC", source_authors[0]).casefold()
+    work_normalized = unicodedata.normalize("NFKC", work_authors[0]).casefold()
+    if source_normalized == work_normalized:
+        return True
     source_parts = _author_identity_parts(source_authors[0])
     work_parts = _author_identity_parts(work_authors[0])
     if len(source_parts) < 2 or len(work_parts) < 2:
@@ -598,9 +605,11 @@ def _leading_author_status(
     work_authors = _dedupe_string_list(work_record.get("authors", []))
     if not source_authors or not work_authors:
         return None
+    source_normalized = unicodedata.normalize("NFKC", source_authors[0]).casefold()
+    work_normalized = unicodedata.normalize("NFKC", work_authors[0]).casefold()
     source_key = _author_key(source_authors[0])
     work_key = _author_key(work_authors[0])
-    status = "match" if source_key and source_key == work_key else "conflict"
+    status = "match" if source_normalized == work_normalized or (source_key and source_key == work_key) else "conflict"
     return {
         "kind": "leading_author",
         "status": status,
@@ -3990,4 +3999,4 @@ def pick_sentences_by_keywords(text: str, keywords: list[str], *, limit: int = 5
         picked.append(normalize_whitespace(sentence))
         if len(picked) >= limit:
             break
-    return picked
+    return picked
\ No newline at end of file