pdf-extraction-fallback-80956b

PDF Extraction Fallback Workflow

This skill provides a robust, multi-layered approach to downloading and extracting text from PDF documents when sources may be protected, corrupted, or inaccessible via standard methods.

When to Use

• Downloading regulatory documents, handbooks, or official PDFs from government/enterprise websites
• Sources that may have JavaScript protection, CORS restrictions, or dynamic content
• When initial PDF downloads produce suspiciously small files or error content
• Any scenario requiring reliable text extraction from potentially problematic PDF sources

Core Workflow

Step 1: Download with Validation

# Download PDF with size check
curl -L -o output.pdf "https://example.com/document.pdf"

# Early failure detection: check file size
file_size=$(stat -c%s output.pdf 2>/dev/null || stat -f%z output.pdf)

if [ "$file_size" -lt 1000 ]; then
    echo "WARNING: File size ($file_size bytes) suggests failed download or error page"
    # Check for HTML/JavaScript error content
    if head -c 500 output.pdf | grep -qi "<html\|<script\|error\|access denied"; then
        echo "FAILURE: File contains error message, not PDF content"
        rm output.pdf
        # Proceed to alternative download method
    fi
fi

Step 2: Sequential Extraction Fallbacks

Try extraction methods in order, moving to next on failure:

Fallback 1: pdftotext (command-line)

if command -v pdftotext &> /dev/null; then
    pdftotext output.pdf output.txt
    if [ -s output.txt ] && [ $(wc -c < output.txt) -gt 100 ]; then
        echo "SUCCESS: pdftotext extraction"
        exit 0
    fi
fi

Fallback 2: PyMuPDF (fitz)

import fitz  # PyMuPDF

def extract_with_pymupdf(pdf_path):
    try:
        doc = fitz.open(pdf_path)
        text = ""
        for page in doc:
            text += page.get_text()
        doc.close()
        if len(text.strip()) > 100:
            return text
        return None
    except Exception as e:
        print(f"PyMuPDF failed: {e}")
        return None