document-skills-pdf

Use pdfplumber to extract text with layout preserved and to detect tables via page.extract_tables(). For plain text, pdftotext from poppler-utils is the fastest option, and results can be converted to pandas DataFrames or Excel files.

First run check_fillable_fields.py to detect form fields. Fillable PDFs are filled with fill_fillable_fields.py using extracted field IDs, while non-fillable PDFs are filled with text annotations via fill_pdf_form_with_annotations.py using bounding box coordinates.

Use pypdf for merging, splitting, rotating, and encryption; pdfplumber for text and table extraction; reportlab for creating new PDFs; and pypdfium2 for fast page rendering. Command-line tools qpdf and pdftotext handle many tasks faster.

Scanned PDFs have no text layer, so standard parsers fail. Convert pages to images with pdf2image, then run OCR with pytesseract to produce searchable text from each page.

ReportLab's built-in fonts lack Unicode subscript and superscript glyphs, so those characters render as solid boxes. Use the <sub> and <super> XML tags inside Paragraph objects instead of Unicode characters.