PDF Processing

PDF Processing Skill

Overview

This skill handles all PDF operations: text extraction, table parsing, form filling, merging, splitting, and metadata extraction.

Step-by-Step Guidance

1. Identify the Operation

Determine what the user needs:

• Text extraction: Read content from a PDF into plain text or markdown
• Table extraction: Pull structured tabular data into CSV/JSON
• Form filling: Write values into fillable PDF form fields
• Merge: Combine multiple PDFs into one document
• Split: Divide a PDF into separate pages or sections
• Metadata: Extract author, creation date, page count, etc.

2. Choose the Right Library

For Python environments:

• pypdf (formerly PyPDF2): text extraction, merging, splitting, metadata - pure Python, no system deps
• pdfplumber: best for tables and layout-aware extraction
• pymupdf (fitz): fastest, handles scanned PDFs with OCR via Tesseract
• reportlab: generate new PDFs or fill forms programmatically
• pdfrw: low-level PDF manipulation, ideal for form filling

3. Text Extraction

import pypdf

reader = pypdf.PdfReader("document.pdf")
text = "\n\n".join(page.extract_text() for page in reader.pages)
print(text)

For layout-preserving extraction:

import pdfplumber

with pdfplumber.open("document.pdf") as pdf:
    for page in pdf.pages:
        print(page.extract_text(layout=True))

4. Table Extraction

import pdfplumber

with pdfplumber.open("document.pdf") as pdf:
    for page in pdf.pages:
        tables = page.extract_tables()
        for table in tables:
            for row in table:
                print(row)

5. Merging PDFs

import pypdf

merger = pypdf.PdfMerger()
for path in ["file1.pdf", "file2.pdf", "file3.pdf"]:
    merger.append(path)
merger.write("merged.pdf")
merger.close()

6. Splitting a PDF

import pypdf

reader = pypdf.PdfReader("document.pdf")
for i, page in enumerate(reader.pages):
    writer = pypdf.PdfWriter()
    writer.add_page(page)
    with open(f"page_{i+1}.pdf", "wb") as f:
        writer.write(f)

7. Form Filling

import pypdf

reader = pypdf.PdfReader("form.pdf")
writer = pypdf.PdfWriter()
writer.append(reader)
writer.update_page_form_field_values(
    writer.pages[0],
    {"field_name": "value", "another_field": "another_value"}
)
with open("filled_form.pdf", "wb") as f:
    writer.write(f)

Edge Cases

• Scanned PDFs (image-only): use pymupdf + Tesseract OCR - warn the user if no text is found
• Password-protected PDFs: call reader.decrypt("password") before reading
• Large PDFs (>100 pages): process page by page to avoid memory issues
• Right-to-left text (Arabic, Hebrew): pdfplumber handles RTL better than pypdf

Output Format

• Plain text extraction → return as a markdown code block
• Table extraction → return as CSV or a markdown table
• File operations (merge/split) → confirm output file path and page count