pdf

PDF Processing

PDF is the universal document format. It is also one of the most annoying formats to work with programmatically. This skill handles the full range of PDF tasks — reading, creating, editing, merging, splitting, and extracting — using the right tool for each job.

Task	Best tool
Extract text (digital PDF)	pdfplumber or pdftotext
Extract tables	pdfplumber
Create PDF from content	reportlab
Merge / split / rotate	pypdf or qpdf
Fill PDF forms	pypdf or pdf-lib (JS)
OCR scanned PDFs	pytesseract + pdf2image
Password protect / decrypt	pypdf or qpdf
Extract images	pdfimages (poppler)

pypdf, pdfplumber, and reportlab are typically pre-installed. Import directly. Only run pip install if an import fails.

---

Reading & Extracting Content

Extract All Text

# pdfplumber — preserves layout better than pypdf
import pdfplumber

with pdfplumber.open("document.pdf") as pdf:
    print(f"Pages: {len(pdf.pages)}")
    for i, page in enumerate(pdf.pages, 1):
        text = page.extract_text()
        if text:
            print(f"\n--- Page {i} ---")
            print(text)

# Command line — fastest for quick extraction
pdftotext document.pdf output.txt         # basic
pdftotext -layout document.pdf output.txt # preserve column layout
pdftotext -f 1 -l 5 document.pdf -        # pages 1–5, stdout

Extract Specific Pages

from pypdf import PdfReader

reader = PdfReader("document.pdf")

# Single page
text = reader.pages[0].extract_text()

# Page range (0-indexed)
for page in reader.pages[2:7]:  # pages 3–7
    print(page.extract_text())

Extract Tables

import pdfplumber
import pandas as pd

with pdfplumber.open("report.pdf") as pdf:
    all_tables = []

    for page_num, page in enumerate(pdf.pages, 1):
        tables = page.extract_tables()

        for table_num, table in enumerate(tables, 1):
            if not table or not table[0]:
                continue

            print(f"Page {page_num}, Table {table_num}: {len(table)} rows")

            # Convert to DataFrame — first row as headers
            df = pd.DataFrame(table[1:], columns=table[0])

            # Clean: strip whitespace, drop empty rows
            df = df.applymap(lambda x: x.strip() if isinstance(x, str) else x)
            df = df.dropna(how="all")

            all_tables.append(df)

# Combine and export
if all_tables:
    combined = pd.concat(all_tables, ignore_index=True)
    combined.to_excel("extracted_tables.xlsx", index=False)
    combined.to_csv("extracted_tables.csv", index=False)
    print(f"Extracted {len(combined)} rows across {len(all_tables)} tables")