working better readings word processor

add-readings.py

@@ -28,39 +28,33 @@ def get_node_reading(node) -> str:
         return katakana_to_hiragana(features[7])
     return node.surface
 
-def annotate_token(surface: str, reading: str) -> str:
+def has_kanji(text: str) -> bool:
-    if not KANJI_RE.search(surface):
+    return bool(KANJI_RE.search(text))
-        return surface
 
-    # full-kanji token like 学校 -> 学校(がっこう)
+def get_node_reading(node) -> str:
-    if all(KANJI_RE.match(ch) for ch in surface):
+    features = node.feature.split(",")
-        return f"{surface}({reading})"
+    ## These depend on what dictionary is being used.
-
+    if len(features) > 6 and features[6] != "*":
-    # prefix kanji + suffix kana, e.g. 行きます / 食べる
+        return katakana_to_hiragana(features[6])
-    m = re.match(r"^([\u4E00-\u9FFF]+)([ぁ-んァ-ンー]*)$", surface)
+    return node.surface
-    if m:
-        kanji_part, kana_suffix = m.groups()
-        hira_suffix = katakana_to_hiragana(kana_suffix)
-        if reading.endswith(hira_suffix):
-            kanji_reading = reading[:len(reading) - len(hira_suffix)] if hira_suffix else reading
-            return f"{kanji_part}({kanji_reading}){kana_suffix}"
-
-    # fallback
-    return f"{surface}({reading})"
 
 def annotate_phrase(text: str) -> str:
     tagger = MeCab.Tagger()
     node = tagger.parseToNode(text)
-
+    vocab = ()
     parts = []
     while node:
         surface = node.surface
         if surface:
+            vocab += (surface,)
+            if has_kanji(surface):
                 reading = get_node_reading(node)
-            parts.append(annotate_token(surface, reading))
+                parts.append(f"{surface}({reading})")
+            else:
+                parts.append(surface)
         node = node.next
 
-    return "".join(parts)
+    return {'reading':("".join(parts)),'vocab':vocab}
 
 #print(annotate_phrase("私は学校へ行きます"))
 
@@ -79,7 +73,11 @@ for cd in dump:
     #     if node.surface:
     #         print(f"{node.surface}\t{node.feature}")
     #     node = node.next
-    print(annotate_phrase(dump[cd]['nm']))
+    anottated = annotate_phrase(dump[cd]['nm'])
+    if anottated['reading'] != dump[cd]['nm']:
+        print(anottated['reading'])
+        dump[cd]['kana'] = anottated['reading']
+        #dump[cd]['vocab'] = anottated['vocab']
     print()
 
 exit()