skills/pptx/scripts/clean.py

"""Remove unreferenced files from an unpacked PPTX directory.

Usage: python clean.py <unpacked_dir>

Example:
    python clean.py unpacked/

This script removes:
- Orphaned slides (not in sldIdLst) and their relationships
- [trash] directory (unreferenced files)
- Orphaned .rels files for deleted resources
- Unreferenced media, embeddings, charts, diagrams, drawings, ink files
- Unreferenced theme files
- Unreferenced notes slides
- Content-Type overrides for deleted files
"""

import re
import sys
from pathlib import Path

import defusedxml.minidom


def get_slides_in_sldidlst(unpacked_dir: Path) -> set[str]:
    pres_path = unpacked_dir / "ppt" / "presentation.xml"
    pres_rels_path = unpacked_dir / "ppt" / "_rels" / "presentation.xml.rels"

    if not pres_path.exists() or not pres_rels_path.exists():
        return set()

    rels_dom = defusedxml.minidom.parse(str(pres_rels_path))
    rid_to_slide = {}
    for rel in rels_dom.getElementsByTagName("Relationship"):
        rid = rel.getAttribute("Id")
        target = rel.getAttribute("Target")
        rel_type = rel.getAttribute("Type")
        if "slide" in rel_type and target.startswith("slides/"):
            rid_to_slide[rid] = target.replace("slides/", "")

    pres_content = pres_path.read_text(encoding="utf-8")
    referenced_rids = set(re.findall(r'<p:sldId[^>]*r:id="([^"]+)"', pres_content))

    return {rid_to_slide[rid] for rid in referenced_rids if rid in rid_to_slide}


def remove_orphaned_slides(unpacked_dir: Path) -> list[str]:
    slides_dir = unpacked_dir / "ppt" / "slides"
    slides_rels_dir = slides_dir / "_rels"
    pres_rels_path = unpacked_dir / "ppt" / "_rels" / "presentation.xml.rels"

    if not slides_dir.exists():
        return []

    referenced_slides = get_slides_in_sldidlst(unpacked_dir)
    removed = []

    for slide_file in slides_dir.glob("slide*.xml"):
        if slide_file.name not in referenced_slides:
            rel_path = slide_file.relative_to(unpacked_dir)
            slide_file.unlink()
            removed.append(str(rel_path))

            rels_file = slides_rels_dir / f"{slide_file.name}.rels"
            if rels_file.exists():
                rels_file.unlink()
                removed.append(str(rels_file.relative_to(unpacked_dir)))

    if removed and pres_rels_path.exists():
        rels_dom = defusedxml.minidom.parse(str(pres_rels_path))
        changed = False

        for rel in list(rels_dom.getElementsByTagName("Relationship")):
            target = rel.getAttribute("Target")
            if target.startswith("slides/"):
                slide_name = target.replace("slides/", "")
                if slide_name not in referenced_slides:
                    if rel.parentNode:
                        rel.parentNode.removeChild(rel)
                        changed = True

        if changed:
            with open(pres_rels_path, "wb") as f:
                f.write(rels_dom.toxml(encoding="utf-8"))

    return removed


def remove_trash_directory(unpacked_dir: Path) -> list[str]:
    trash_dir = unpacked_dir / "[trash]"
    removed = []

    if trash_dir.exists() and trash_dir.is_dir():
        for file_path in trash_dir.iterdir():
            if file_path.is_file():
                rel_path = file_path.relative_to(unpacked_dir)
                removed.append(str(rel_path))
                file_path.unlink()
        trash_dir.rmdir()

    return removed


def get_slide_referenced_files(unpacked_dir: Path) -> set:
    referenced = set()
    slides_rels_dir = unpacked_dir / "ppt" / "slides" / "_rels"

    if not slides_rels_dir.exists():
        return referenced

    for rels_file in slides_rels_dir.glob("*.rels"):
        dom = defusedxml.minidom.parse(str(rels_file))
        for rel in dom.getElementsByTagName("Relationship"):
            target = rel.getAttribute("Target")
            if not target:
                continue
            target_path = (rels_file.parent.parent / target).resolve()
            try:
                referenced.add(target_path.relative_to(unpacked_dir.resolve()))
            except ValueError:
                pass

    return referenced


def remove_orphaned_rels_files(unpacked_dir: Path) -> list[str]:
    resource_dirs = ["charts", "diagrams", "drawings"]
    removed = []
    slide_referenced = get_slide_referenced_files(unpacked_dir)

    for dir_name in resource_dirs:
        rels_dir = unpacked_dir / "ppt" / dir_name / "_rels"
        if not rels_dir.exists():
            continue

        for rels_file in rels_dir.glob("*.rels"):
            resource_file = rels_dir.parent / rels_file.name.replace(".rels", "")
            try:
                resource_rel_path = resource_file.resolve().relative_to(unpacked_dir.resolve())
            except ValueError:
                continue

            if not resource_file.exists() or resource_rel_path not in slide_referenced:
                rels_file.unlink()
                rel_path = rels_file.relative_to(unpacked_dir)
                removed.append(str(rel_path))

    return removed


def get_referenced_files(unpacked_dir: Path) -> set:
    referenced = set()

    for rels_file
…