06_redaction.py

"""
06 - Redaction.

Three redaction modes:
- Text-based (literal string match)
- Regex (e.g. SSN patterns)
- Area-based (rectangle on a specific page)

Redaction is destructive - the underlying text is removed from the content
stream, not just visually masked. This makes it suitable for HIPAA, GDPR,
and similar compliance requirements.

Run:
    python examples/06_redaction.py
"""

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

import exis_pdfeditor


def main() -> None:
    init_license()
    header("Demo 06 - Redaction")

    sample = get_sample_pdf()
    out = ensure_output_dir()

    # Text + regex redaction
    info("Text + regex redaction...")
    result = exis_pdfeditor.redact(
        str(sample),
        str(out / "06-text-redacted.pdf"),
        redactions=[
            {"text": "the", "replaceWith": "[X]", "caseSensitive": False},
            {"text": r"\d{3}-\d{2}-\d{4}", "isRegex": True, "replaceWith": "[SSN]"},
            {"text": r"[\w\.-]+@[\w\.-]+\.\w+", "isRegex": True, "replaceWith": "[EMAIL]"},
        ],
    )
    success(f"Applied {result.redactionsApplied} redaction(s)")

    # Area-based: black out a 200x20pt rectangle on page 1
    info("Area-based redaction (rectangle on page 1)...")
    area_result = exis_pdfeditor.redact(
        str(sample),
        str(out / "06-area-redacted.pdf"),
        redactions=[
            {
                "area": {"x": 100, "y": 700, "width": 200, "height": 20},
                "pageNumber": 1,
            },
        ],
    )
    success(f"Area redactions applied: {area_result.redactionsApplied}")

    print()
    info(f"Output: {out}")


if __name__ == "__main__":
    main()