| |
| """One-command Arabic scanned document conversion to editable RTL Word.""" |
| import argparse, json, subprocess, tempfile |
| from pathlib import Path |
|
|
| ROOT = Path(__file__).resolve().parent |
| BUNDLED_PYTHON = Path("/Users/abdulrahmanalshehry/.cache/codex-runtimes/codex-primary-runtime/dependencies/python/bin/python3") |
|
|
| def main(): |
| ap = argparse.ArgumentParser(description="Convert Arabic images, scanned PDFs, or scanned Word files to RTL DOCX") |
| ap.add_argument("input", type=Path); ap.add_argument("output", type=Path) |
| ap.add_argument("--dpi", type=int, default=150); ap.add_argument("--resume", action="store_true"); ap.add_argument("--layout", action="store_true", help="Use Abjad vision layout mode for tables and structured pages"); ap.add_argument("--trained", action="store_true", help="Use the trained MLX LoRA adapter instead of fast macOS Vision") |
| args = ap.parse_args() |
| with tempfile.TemporaryDirectory(prefix="abjad-convert-") as td: |
| temp = Path(td); text = temp / "ocr.txt" |
| venv_python = ROOT / ".venv" / "bin" / "python" |
| word_python = str(BUNDLED_PYTHON) if BUNDLED_PYTHON.exists() else "python3" |
| python = str(venv_python) if args.trained and venv_python.exists() else word_python |
| if args.layout and args.trained: |
| ap.error("--layout and --trained are mutually exclusive") |
| if args.layout: |
| subprocess.run([python, str(ROOT / "extract_to_word.py"), str(args.input), "-o", str(args.output), "--model", "Abjad_0.1", "--dpi", str(args.dpi)], check=True) |
| print(json.dumps({"word_output": str(args.output), "mode": "layout"}, ensure_ascii=False)); return |
| engine = "macOS Vision Arabic OCR" |
| pipeline = ROOT / "vision_pipeline.py" |
| if args.trained: |
| pipeline = ROOT / "mlx_ocr_pipeline.py" |
| engine = "Abjad MLX LoRA Arabic OCR" |
| command = [python, str(pipeline), str(args.input), "-o", str(text), "--dpi", str(args.dpi)] |
| if args.trained: |
| command.extend(["--adapter", str(ROOT / "adapters")]) |
| if args.resume: command.append("--resume") |
| subprocess.run(command, check=True) |
| subprocess.run([word_python, str(ROOT / "text_to_word.py"), str(text), str(args.output)], check=True) |
| report = json.loads((text.with_suffix(".progress.json")).read_text(encoding="utf-8")) |
| report["word_output"] = str(args.output) |
| args.output.with_suffix(".quality.json").write_text(json.dumps(report, ensure_ascii=False, indent=2), encoding="utf-8") |
| print(json.dumps(report, ensure_ascii=False, indent=2)) |
|
|
| if __name__ == "__main__": main() |
|
|
