62 lines
2.5 KiB
Python
62 lines
2.5 KiB
Python
#!/usr/bin/env python3
|
|
"""
|
|
doocus tree indexer — replicate a local drive tree into docs-output/index.json.
|
|
|
|
Walks a downloaded-drive root, preserving the full folder hierarchy. Locally
|
|
extracts text for docx/pdf/pptx/xlsx (search-index sidecars, never a replacement
|
|
for the original), flags videos as meetus meetings, and links everything else to
|
|
the original without duplicating it. Deterministic and offline.
|
|
|
|
uv run process_tree.py /mnt/drive # → docs-output/index.json
|
|
uv run process_tree.py /mnt/drive --output out/docs
|
|
uv run process_tree.py /mnt/drive --dry-run # counts only, no writes
|
|
"""
|
|
import argparse
|
|
import logging
|
|
import sys
|
|
|
|
from doocus.tree import build_index
|
|
|
|
|
|
def main() -> int:
|
|
parser = argparse.ArgumentParser(description="Index a local drive tree for doocus.")
|
|
parser.add_argument("source", help="Local drive-download root to index")
|
|
parser.add_argument("--output", default="docs-output",
|
|
help="Output directory for index.json + sidecars (default: docs-output)")
|
|
parser.add_argument("--render", action="store_true",
|
|
help="Render page/slide images where supported (needs extra deps)")
|
|
parser.add_argument("--ocr", action="store_true",
|
|
help="OCR scanned PDFs where supported")
|
|
parser.add_argument("--dry-run", action="store_true",
|
|
help="Walk and classify only; write nothing")
|
|
parser.add_argument("--verbose", action="store_true", help="Verbose logging")
|
|
args = parser.parse_args()
|
|
|
|
logging.basicConfig(level=logging.DEBUG if args.verbose else logging.INFO,
|
|
format="%(message)s")
|
|
|
|
try:
|
|
index = build_index(
|
|
args.source,
|
|
args.output,
|
|
options={"render": args.render, "ocr": args.ocr},
|
|
dry_run=args.dry_run,
|
|
)
|
|
except NotADirectoryError as e:
|
|
print(f"ERROR: {e}", file=sys.stderr)
|
|
return 1
|
|
|
|
c = index["counts"]
|
|
print(f"\n{c['total']} files: {c['extracted']} extracted, "
|
|
f"{c['meeting']} meeting, {c['link']} linked"
|
|
+ (f", {c['warnings']} with warnings" if c.get("warnings") else "")
|
|
+ (" (dry run — nothing written)" if args.dry_run else ""))
|
|
if not args.dry_run and c["meeting"]:
|
|
print(f"→ {args.output}/meetings.txt lists the {c['meeting']} meeting path(s) "
|
|
f"for meetus: `ctrl/batch.sh -i <mount> -o {args.output} -l {args.output}/meetings.txt`")
|
|
return 0
|
|
|
|
|
|
if __name__ == "__main__":
|
|
sys.exit(main())
|