INNER CODE UNIT · Python
parse_elements
kangarooking/cangjie-skill · scripts/build_chunks.py:38
def parse_elements(text: str) -> list[dict]:
"""把 Markdown/TXT 解析为 SourceDocument elements（标题/段落/列表/代码）。"""
elements: list[dict] = []
heading_stack: list[tuple[int, str]] = []
in_code = False
buf: list[str] = []
buf_type = "paragraph"
def flush():
nonlocal buf, buf_type
content = "\n".join(buf).strip()
buf = []
if not content:
return
elements.append({
"element_id": f"el-{len(elements):06d}",
"type": buf_type,
"text": content,