Removing headers/footers before text extraction. Pattern #7: Layout-Preserving Text Extraction (pdfplumber) The Impact: PyMuPDF extracts raw text, but pdfplumber excels at preserving column layout and reading multi-column scientific papers.
import subprocess def ocr_pdf_powerful(input_pdf: str, output_pdf: str, language="eng"): cmd = [ "ocrmypdf", "--language", language, "--deskew", "--clean", "--pdfa-image-compression", "jpeg", input_pdf, output_pdf ] subprocess.run(cmd, check=True) Removing headers/footers before text extraction
CSS for print media ( @media print ) ensures pixel-perfect rendering. Pattern #10: Adding Digital Signatures (Modern Compliance) The Impact: eIDAS, ESIGN, and 21 CFR Part 11 require cryptographic signatures. PyMuPDF 1.23+ supports PKCS#7 signatures. Pattern #4: PDF to Image Conversion (for ML
def redact_sensitive_text(pdf_path: str, output_path: str, search_terms: list): doc = fitz.open(pdf_path) for page in doc: for term in search_terms: text_instances = page.search_for(term) for inst in text_instances: page.add_redact_annot(inst, fill=(0,0,0)) # black redaction page.apply_redactions() doc.save(output_path) doc.close() Add metadata tracking which redactions occurred (audit log). Pattern #4: PDF to Image Conversion (for ML Pipelines) The Impact: PDFs feed vision models. Convert to PNG/JPEG at 300+ DPI without losing vector quality. language="eng"): cmd = [ "ocrmypdf"
Parallelize across pages using concurrent.futures for PDFs over 500 pages. Pattern #2: Vector-Accurate Table Extraction (Better than Tabula) The Impact: PDF tables are not true data structures. Using PyMuPDF’s get_text("words") with geometric clustering yields verified 99% accuracy.
Use add_redact_annot() followed by apply_redactions() .