pdf

Use pdfplumber's page.extract_tables() method to detect table structures and return cell data, then convert results to a pandas DataFrame. For poorly formatted tables, crop the page to a bounding box before extraction.

Use pypdf's PdfMerger class: append each PDF path in order, then call write() with the output path. This preserves page order, which matters for submissions requiring control sheet first, then paper, then appendix.

pypdf handles basic text extraction, merging, and metadata, while pdfplumber preserves layout better and excels at table extraction. PyMuPDF (fitz) is preferred for rendering page regions as images, such as equations.

Standard parsers cannot read scanned image PDFs. Convert pages to images with pdf2image, then run pytesseract OCR on each image to recover machine-readable text.

Open the file with pypdf's PdfReader and count len(reader.pages), subtracting excluded pages like the control sheet or appendix. Compare the result against the limit, such as 25 pages for MCM main papers.

PDFs often store equations as images or fragmented text glyphs, so text extraction fails. The reliable approach is rendering the equation region as an image with PyMuPDF and manually transcribing it to LaTeX.