pdf

Core Rules

1. Deterministic AcroForm Discovery & Population: Inspect fillable PDF forms using pypdf (reader.get_fields()). Programmatically populate form widgets with strict field-name mapping before attempting visual raster stamping.
2. Tripartite PDF Extraction: Combine pypdf for fast text/metadata extraction, pdfplumber for tabular cell coordinate extraction, and tesseract/OCR for scanned raster pages.
3. Visual Layout Verification: Never assume programmatic PDF generation rendered cleanly without visual checks. Convert pages to PNG via pdftoppm -png -r 150 doc.pdf page to inspect table border alignment, font clipping, and multi-page overflows.
4. Signature & Bounding Box Integrity: When processing proofs of delivery (ePOD) or legal contracts, verify signature vector presence ($w \ge 100\text{px}, h \ge 50\text{px}$) and timestamps. Never accept blank or 0-byte signature fields.
5. Font Embedding & ASCII Dash Rule: Embedded fonts must be subsetted. Hyphens must use standard ASCII (-, U+002D) rather than non-breaking hyphens (U+2011) which cause corrupted glyph rendering in headless PDF engines.

Suggested Process