v0.5.4

NewPromt
- small fix

word2md/walker.py

@@ -36,8 +36,8 @@ from .classify import (
     is_code_style,
     is_special_title,
     is_toc_style,
-    look_like_list_item,
     parse_biblio_line,
+    parse_list_item,
     paragraph_has_numpr,
     style_name,
 )
@@ -182,6 +182,7 @@ class DocxWalker:
         self._code_caption: CaptionBlock | None = None
         self._list_buf: list[str] | None = None
         self._list_ordered: bool | None = None
+        self._list_levels: list[int] | None = None
 
     def walk(self) -> list[Block]:
         blocks: list[Block] = []
@@ -225,10 +226,12 @@ class DocxWalker:
             ListBlock(
                 ordered=bool(self._list_ordered),
                 items=list(self._list_buf),
+                item_levels=list(self._list_levels or []),
             )
         )
         self._list_buf = None
         self._list_ordered = None
+        self._list_levels = None
 
     def _handle_paragraph(self, paragraph: DocxParagraph) -> list[Block]:
         out: list[Block] = []
@@ -408,33 +411,37 @@ class DocxWalker:
             return None
         self._flush_code(out)
         # python-docx does not expose the rendered number, so infer from the text
-        m = re.match(r"^\d+\.\s*(.*)$", para.text)
-        if m:
-            ordered, body = True, m.group(1)
+        parsed = parse_list_item(para.text)
+        if parsed is not None:
+            ordered, level, body = parsed
         else:
-            ordered, body = False, re.sub(r"^[–—\-•▪]\s*", "", para.text)
-        self._append_list_item(ordered, body, out)
+            ordered, level, body = False, 1, re.sub(r"^[–—\-•▪]\s*", "", para.text)
+        self._append_list_item(ordered, body, out, level)
         return []
 
     def _as_marker_list_item(self, para: _Para, out: list[Block]) -> list[Block] | None:
-        list_hit = look_like_list_item(para.text)
+        parsed = parse_list_item(para.text)
         indented = para.paragraph.paragraph_format.left_indent or "\t" in (
             para.paragraph.text or ""
         )
-        if not (list_hit and indented):
+        if not (parsed and indented):
             return None
-        ordered, _ok, body = list_hit
+        ordered, level, body = parsed
         self._flush_code(out)
-        self._append_list_item(ordered, body, out)
+        self._append_list_item(ordered, body, out, level)
         return []
 
-    def _append_list_item(self, ordered: bool, body: str, out: list[Block]) -> None:
+    def _append_list_item(
+        self, ordered: bool, body: str, out: list[Block], level: int = 1
+    ) -> None:
         if self._list_buf is not None and self._list_ordered == ordered:
             self._list_buf.append(body)
+            self._list_levels.append(level)
             return
         self._flush_list(out)
         self._list_buf = [body]
         self._list_ordered = ordered
+        self._list_levels = [level]
 
     def _as_plain_paragraph(self, para: _Para, out: list[Block]) -> list[Block]:
         self._flush_code(out)