INNER CODE UNIT · Python

_leading_number

Laurent00TT/PharosRAG · src/chunker/core.py:142

def _leading_number(txt):
    """Parse a leading section/list number -> ('dotted', depth) | ('bare', value) | None.
    Year-leading (2026 ...) is treated as not-a-number (review F1/F4)."""
    m = NUM_RE.match(txt)
    if not m:
        return None
    tok = m.group(1)
    if YEAR_RE.match(tok.split(".")[0]):
        return None
    if "." in tok:
        return ("dotted", tok.count(".") + 1)
    if tok.isdigit():
        return ("bare", int(tok))
    return None


def heading_level(el: Element, doc_type, promote_bare=True):
    """text_level primary (free from parser). DOTTED numbering (2.1) always refines depth;

View source record →

📰 Research Paper
Loading…
⏳ Fetching content…