pdf-processor

PDF Processor Skill

Overview

The PDF Processor skill provides comprehensive PDF document handling capabilities including extraction, analysis, manipulation, and generation of PDF files for documentation, reporting, and compliance purposes.

Features

1. Text Extraction

• Extract plain text from PDFs
• Preserve formatting and structure
• Extract tables and structured data
• Multi-language support
• Handle encrypted PDFs

2. OCR Processing

• Convert scanned documents to text
• Support for 100+ languages
• Image preprocessing for better accuracy
• Handwriting recognition
• Layout analysis

3. Metadata Operations

• Extract document properties
• Read/write custom metadata
• Extract embedded files
• Digital signature verification
• Creation/modification date tracking

4. PDF Manipulation

• Merge multiple PDFs
• Split PDFs by pages or bookmarks
• Rotate pages
• Crop and resize
• Add watermarks and stamps

5. Form Processing

• Extract form fields
• Fill PDF forms programmatically
• Validate form data
• Create fillable forms
• Export form data to JSON/CSV

6. Report Generation

• Generate PDFs from templates
• Create reports from data
• Add charts and graphs
• Include images and logos
• Apply corporate branding

Configuration

{
  "pdf_processor": {
    "enabled": true,
    "ocr": {
      "enabled": true,
      "languages": ["eng", "fra", "deu", "spa"],
      "dpi": 300,
      "preprocessing": true
    },
    "extraction": {
      "preserve_formatting": true,
      "extract_images": true,
      "extract_tables": true,
      "extract_metadata": true
    },
    "security": {
      "allow_encrypted": true,
      "max_file_size_mb": 100,
      "sandbox_mode": true
    },
    "output": {
      "formats": ["text", "json", "html", "markdown"],
      "compression": true,
      "optimization": true
    },
    "performance": {
      "parallel_processing": true,
      "max_workers": 4,
      "cache_enabled": true
    }
  }
}

Usage Examples

Text Extraction

# Extract text from a PDF
from pdf_processor import PDFExtractor

extractor = PDFExtractor()
text = extractor.extract_text('document.pdf')

# Extract with formatting preserved
formatted_text = extractor.extract_text(
    'document.pdf',
    preserve_formatting=True
)

# Extract specific pages
page_text = extractor.extract_pages(
    'document.pdf',
    pages=[1, 3, 5]
)

OCR Processing

# Process scanned PDF with OCR
from pdf_processor import OCRProcessor

ocr = OCRProcessor(languages=['eng', 'spa'])
text = ocr.process_scanned_pdf('scanned.pdf')

# With preprocessing for better accuracy
text = ocr.process_scanned_pdf(
    'scanned.pdf',
    preprocess=True,
    deskew=True,
    denoise=True
)

Table Extraction

# Extract tables from PDF
from pdf_processor import TableExtractor

extractor = TableExtractor()
tables = extractor.extract_tables('report.pdf')

for idx, table in enumerate(tables):
    # Convert to pandas DataFrame
    df = table.to_dataframe()
    # Export to CSV
    df.to_csv(f'table_{idx}.csv')

PDF Generation

# Generate PDF report from template
from pdf_processor import ReportGenerator

generator = ReportGenerator()

data = {
    'title': 'Monthly DevOps Report',
    'date': '2024-01-15',
    'metrics': {
        'uptime': '99.9%',
        'deployments': 47,
        'incidents': 2
    },
    'charts': ['uptime_chart.png', 'deployment_trend.png']
}

generator.create_report(
    template='monthly_report_template.html',
    data=data,
    output='monthly_report.pdf'
)

Form Processing

# Extract and fill PDF forms
from pdf_processor import FormProcessor

processor = FormProcessor()

# Extract form fields
fields = processor.extract_fields('form.pdf')
print(f"Found {len(fields)} form fields")

# Fill form with data
form_data = {
    'name': 'John Doe',
    'email': '[email protected]',
    'department': 'Engineering'
}

processor.fill_form(
    'form.pdf',
    form_data,
    output='filled_form.pdf'
)

PDF Manipulation

# Merge multiple PDFs
from pdf_processor import PDFManipulator

manipulator = PDFManipulator()

# Merge PDFs
manipulator.merge_pdfs(
    ['doc1.pdf', 'doc2.pdf', 'doc3.pdf'],
    output='merged.pdf'
)

# Split PDF by pages
manipulator.split_pdf(
    'large_document.pdf',
    pages_per_file=10,
    output_dir='split_docs/'
)

# Add watermark
manipulator.add_watermark(
    'document.pdf',
    watermark='CONFIDENTIAL',
    output='watermarked.pdf',
    opacity=0.3
)

Integration Examples

Compliance Report Generation

# Generate compliance reports from audit data
def generate_compliance_report(audit_data):
    generator = ReportGenerator()
    
    # Create PDF with audit findings
    report = generator.create_report(
        template='compliance_template.pdf',
        data={
            'audit_date': audit_data['date'],
            'findings': audit_data['findings'],
            'recommendations': audit_data['recommendations'],
            'compliance_score': audit_data['score']
        }
    )
    
    # Add digital signature
    report.sign(
        certificate='company_cert.p12',
        password='cert_password'
    )
    
    return report

Documentation Processing Pipeline

# Process technical documentation
class DocProcessor:
    def process_documentation(self, pdf_path):
        # Extract text and metadata
        text = self.extract_text(pdf_path)
        metadata = self.extract_metadata(pdf_path)
        
        # Extract code snippets
        code_blocks = self.extract_code_blocks(text)
        
        # Extract diagrams and charts
        images = self.extract_images(pdf_path)
        
        # Generate searchable index
        index = self.create_search_index(text)
        
        # Convert to multiple formats
        self.export_to_markdown(text, 'docs.md')
        self.export_to_html(text, 'docs.html')
        
        return {
            'text': text,
            'metadata': metadata,
            'code_blocks': code_blocks,
            'images': images,
            'index': index
        }

Advanced Features

Batch Processing

# Process multiple PDFs in parallel
from pdf_processor import BatchProcessor

processor = BatchProcessor(max_workers=4)

# Define processing pipeline
pipeline = [
    ('extract_text', {}),
    ('extract_tables', {}),
    ('extract_metadata', {})
]

# Process all PDFs in directory
results = processor.process_directory(
    'documents/',
    pipeline=pipeline,
    output_format='json'
)

Intelligent Data Extraction

# Extract specific data using patterns
from pdf_processor import IntelligentExtractor

extractor = IntelligentExtractor()

# Define extraction patterns
patterns = {
    'invoice_number': r'Invoice #: (\d+)',
    'total_amount': r'Total: \$([\d,]+\.\d{2})',
    'date': r'Date: (\d{2}/\d{2}/\d{4})'
}

# Extract structured data
data = extractor.extract_by_patterns(
    'invoice.pdf',
    patterns=patterns
)

Performance Optimization

Caching Strategy

# Enable caching for repeated operations
from pdf_processor import CachedProcessor

processor = CachedProcessor(
    cache_dir='/tmp/pdf_cache',
    ttl=3600  # Cache for 1 hour
)

# Subsequent calls use cache
text1 = processor.extract_text('large_doc.pdf')  # Slow
text2 = processor.extract_text('large_doc.pdf')  # Fast (cached)

Memory Management

# Stream processing for large PDFs
from pdf_processor import StreamProcessor

processor = StreamProcessor()

# Process large PDF in chunks
for chunk in processor.stream_pages('huge_document.pdf', chunk_size=10):
    # Process 10 pages at a time
    process_chunk(chunk)

Error Handling

# Robust error handling
from pdf_processor import PDFProcessor, PDFError

try:
    processor = PDFProcessor()
    result = processor.process('document.pdf')
except PDFError.CorruptedFile as e:
    print(f"PDF is corrupted: {e}")
    # Attempt repair
    repaired = processor.repair_pdf('document.pdf')
except PDFError.PasswordProtected as e:
    print(f"PDF is password protected")
    # Request password
    password = input("Enter PDF password: ")
    result = processor.process('document.pdf', password=password)
except PDFError.UnsupportedFormat as e:
    print(f"Unsupported PDF format: {e}")

Scripts Directory

The PDF processor includes utility scripts in the scripts/ directory:

pdf-extract.py

#!/usr/bin/env python3
# Extract text from PDFs via command line

import argparse
from pdf_processor import PDFExtractor

def main():
    parser = argparse.ArgumentParser()
    parser.add_argument('input', help='Input PDF file')
    parser.add_argument('-o', '--output', help='Output file')
    parser.add_argument('--format', choices=['text', 'json', 'html'],
                       default='text')
    args = parser.parse_args()
    
    extractor = PDFExtractor()
    result = extractor.extract(args.input, format=args.format)
    
    if args.output:
        with open(args.output, 'w') as f:
            f.write(result)
    else:
        print(result)

if __name__ == '__main__':
    main()

pdf-merge.sh

#!/bin/bash
# Merge multiple PDFs

if [ $# -lt 2 ]; then
    echo "Usage: $0 output.pdf input1.pdf input2.pdf ..."
    exit 1
fi

OUTPUT=$1
shift

python3 -c "
from pdf_processor import PDFManipulator
m = PDFManipulator()
m.merge_pdfs(['$@'], '$OUTPUT')
print(f'Merged {len(['$@'])} PDFs into $OUTPUT')
"

Troubleshooting

Common Issues

1. OCR Accuracy Issues

   • Solution: Increase DPI, enable preprocessing
   • Check language settings

2. Memory Issues with Large PDFs

   • Solution: Use streaming mode
   • Process in chunks

3. Corrupted PDF Files

   • Solution: Use repair function
   • Try alternative extraction methods

4. Missing Dependencies

   • Install: pip install pypdf2 pdfplumber pytesseract
   • Install system deps: apt-get install tesseract-ocr poppler-utils

Best Practices

1. Always validate input PDFs
2. Use appropriate error handling
3. Enable caching for repeated operations
4. Stream large files instead of loading into memory
5. Sanitize user-uploaded PDFs
6. Respect PDF permissions and DRM
7. Optimize PDFs after manipulation
8. Use async processing for web applications