contraction-fix 0.1.5

A fast and efficient library for fixing contractions in text

Contraction Fix

A fast and efficient library for fixing contractions in text. This package provides tools to expand contractions in English text while maintaining high performance and accuracy.

Features

• Fast text processing using precompiled regex patterns
• Support for standard contractions, informal contractions, and internet slang
• Configurable dictionary usage
• Caching for improved performance
• Preview functionality to see contractions before fixing
• Easy addition and removal of custom contractions

Installation

pip install contraction-fix

Usage

Basic Usage

from contraction_fix import fix

text = "I can't believe it's not butter!"
fixed_text = fix(text)
print(fixed_text)  # "I cannot believe it is not butter!"

Advanced Usage

from contraction_fix import ContractionFixer

# Create a custom fixer instance
fixer = ContractionFixer(use_informal=True, use_slang=False)

# Fix text
text = "I'd like to see y'all tomorrow"
fixed_text = fixer.fix(text)
print(fixed_text)  # "I would like to see you all tomorrow"

# Preview contractions
matches = fixer.preview(text, context_size=5)
for match in matches:
    print(f"Found '{match['match']}' at position {match['start']}")
    print(f"Context: '{match['context']}'")
    print(f"Will be replaced with: '{match['replacement']}'")

# Add custom contraction
fixer.add_contraction("gonna", "going to")

# Remove contraction
fixer.remove_contraction("won't")

Dictionary Types

The package uses three types of dictionaries:

1. Standard Contractions: Common English contractions like "can't", "won't", etc.
2. Informal Contractions: Less formal contractions and patterns like "goin'", "doin'", etc.
3. Internet Slang: Modern internet slang and abbreviations like "lol", "btw", etc.

Performance

The package is optimized for speed through:

• Precompiled regex patterns
• LRU caching of results
• Efficient dictionary lookups
• Minimal memory usage

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.