phuetz/code-buddy/doc-ingest

python3 - "$FILE" <<'PY'
import sys, os
src = sys.argv[1]
ext = os.path.splitext(src)[1].lower()
 
def pdf(p):
    from pypdf import PdfReader
    return "\n\n".join((pg.extract_text() or "") for pg in PdfReader(p).pages)
 
def docx(p):
    import docx
    d = docx.Document(p)
    return "\n".join(par.text for par in d.paragraphs)
 
def pptx(p):
    from pptx import Presentation
    out = []
    for i, sl in enumerate(Presentation(p).slides, 1):
        out.append(f"## Slide {i}")
        for sh in sl.shapes:
            if sh.has_text_frame:
                out.append(sh.text_frame.text)
    return "\n".join(out)
 
def xlsx(p):
    from openpyxl import load_workbook
    wb = load_workbook(p, data_only=True)
    out = []
    for ws in wb.worksheets:
        out.append(f"## {ws.title}")
        for row in ws.iter_rows(values_only=True):
            out.append(" | ".join("" if c is None else str(c) for c in row))
    return "\n".join(out)
 
handlers = {".pdf": pdf, ".docx": docx, ".pptx": pptx, ".xlsx": xlsx}
fn = handlers.get(ext)
if not fn:
    sys.exit(f"Lean path can't read {ext}; enable the rich backend (prepare:python:extras) for markitdown.")
text = fn(src)
assert text and text.strip(), "extraction produced no text"
print(f"[{len(text)} chars]")
print("\n".join(text.splitlines()[:40]))
PY

python3 - <<'PY'
try:
    from markitdown import MarkItDown
except ModuleNotFoundError:
    raise SystemExit(
        "markitdown not installed. Run `npm run prepare:python:extras` in cowork/ "
        "to add the optional ingestion backend (HTML, images/OCR, EPUB, ...)."
    )
md = MarkItDown()
res = md.convert("INPUT_PATH")
assert res.text_content, "markitdown produced no content"
print(res.text_content[:4000])
PY