Function find_orphans

openkb/lint.py:292–345

Find pages that have no links to or from other pages. A page is orphaned if: - No other page links to it (no incoming links), AND - It has no outgoing wikilinks itself. index.md is excluded from orphan detection. Args: wiki: Path to the wiki root directory. Return

(wiki: Path)

def find_orphans(wiki: Path) -> list[str]:
    """Find pages that have no links to or from other pages.

    A page is orphaned if:
    - No other page links to it (no incoming links), AND
    - It has no outgoing wikilinks itself.

    index.md is excluded from orphan detection.

    Args:
        wiki: Path to the wiki root directory.

    Returns:
        List of relative page paths that are orphaned.
    """
    # Exclude index, schema, log, and sources/ (sources are auto-generated, not expected to be linked)
    all_mds = [
        p
        for p in wiki.rglob("*.md")
        if p.name not in {"index.md", *_EXCLUDED_FILES}
        and "sources" not in p.relative_to(wiki).parts
    ]
    if not all_mds:
        return []

    # Build outgoing links per page
    outgoing: dict[str, set[str]] = {}
    for md in all_mds:
        rel = str(md.relative_to(wiki).with_suffix("")).replace("\\", "/")
        text = _read_md(md)
        outgoing[rel] = set(_extract_wikilinks(text))

    # Build incoming link set (which pages are linked to)
    incoming: set[str] = set()
    for links in outgoing.values():
        for lnk in links:
            cleaned = lnk.strip().strip("/")
            incoming.add(cleaned)
            # Only treat a link as a bare-stem reference when it is itself a
            # bare stem. A qualified link like ``concepts/foo`` must not mark a
            # same-stem page in another directory (``summaries/foo``) as linked,
            # which would hide a genuine orphan.
            if "/" not in cleaned:
                incoming.add(Path(cleaned).stem)

    orphans: list[str] = []
    for rel, links in outgoing.items():
        stem = Path(rel).stem
        has_incoming = rel in incoming or stem in incoming
        has_outgoing = bool(links)
        if not has_incoming and not has_outgoing:
            orphans.append(rel)

    return sorted(orphans)


def find_missing_entries(