19_ocr.py

"""
19 - OCR for scanned PDFs (Windows wheels only).

- make_searchable_pdf: add an invisible text layer to scanned pages
- redact_scanned_pdf: find terms via OCR and redact them

On Linux/macOS this script catches OcrNotSupportedError and exits cleanly.

Run:
    python examples/19_ocr.py
"""

from _common import (
    header, info, success, warn, ensure_output_dir, get_sample_pdf, init_license,
)

import exis_pdfeditor
from exis_pdfeditor import OcrNotSupportedError


def main() -> None:
    init_license()
    header("Demo 19 - OCR (Windows only)")

    sample = get_sample_pdf()
    out = ensure_output_dir()

    try:
        info("make_searchable_pdf...")
        result = exis_pdfeditor.make_searchable_pdf(
            str(sample),
            str(out / "19-searchable.pdf"),
            languages=["eng"],
            dpi=300,
        )
        success(f"make_searchable_pdf -> {result}")

        info("redact_scanned_pdf (sample terms)...")
        result = exis_pdfeditor.redact_scanned_pdf(
            str(sample),
            str(out / "19-redact-scanned.pdf"),
            terms=["CONFIDENTIAL", "SSN"],
            visible_replacement="[REDACTED]",
        )
        success(f"redact_scanned_pdf -> {result}")
    except OcrNotSupportedError as ex:
        warn(f"OCR not available on this platform/wheel: {ex}")
        info("Install the Windows wheel to exercise OCR demos.")


if __name__ == "__main__":
    main()