nlpprepkit 1.2.0

Text Preprocessing Library

nlpprepkit

This can't be the best library for text preprocessing, but it's definitely a library!

Installation

pip install nlpprepkit

Or install from source:

git clone https://github.com/vnniciusg/nlpprepkit.git
cd nlpprepkit
pip install -e .

Features

• Flexible cleaning options: Control which cleaning operations to apply
• Parallel processing: Process large text collections efficiently with multi-core support
• Caching: Avoid redundant processing with built-in caching system
• NLTK integration: Easy access to stemming, lemmatization, and stopword removal
• Configurable: Save and load configuration settings for reproducible workflows

Quick Start

from nlpprepkit import TextPreprocessor

# Create a preprocessor with default settings
preprocessor = TextPreprocessor()

# Process a single text
cleaned_text = preprocessor.process_text("Check out this URL: https://example.com and these numbers 12345!")
print(cleaned_text) # Output: "check url number"

# Process multiple texts in parallel
texts = [
    "First text with URL https://example.org",
    "Second text with numbers 12345",
    "Third text with emoji 😀 and contraction don't"
]
results = preprocessor.process_text(texts)

Customizing Configuration

from nlpprepkit import TextPreprocessor, CleaningConfig

# Create a custom configuration
config = CleaningConfig(
    expand_contractions=True,
    lowercase=True,
    remove_urls=True,
    remove_newlines=True,
    remove_numbers=True,
    remove_punctuation=True,
    remove_emojis=True,
    tokenize=True,
    remove_stopwords=True,
    stemming=False,
    lemmatization=True, # Only one of stemming or lemmatization can be enabled
    normalize_unicode=True,
    language="english",
    custom_stopwords=["custom1", "custom2"],
    keep_words=["important1", "important2"],
    min_word_length=2,
    max_word_length=15
)

# Create preprocessor with custom config
preprocessor = TextPreprocessor(config)

Saving and Loading Configuration

# Save configuration to a file
preprocessor.save_config("my_config.json")

# Load configuration from a file
preprocessor = TextPreprocessor.from_config_file("my_config.json")

Caching

The library includes a caching system to avoid redundant processing:

# Enable caching (enabled by default)
TextPreprocessor.enable_cache(max_size=1000)

# Clear cache if needed
TextPreprocessor.clear_cache()

Parallel Processing

# Process a large list of texts with parallel processing
results = preprocessor.process_text(
    large_text_list,
    max_workers=8, # Number of parallel workers
    batch_size=1000 # Batch size for processing
)

Available Cleaning Operations

• Expand contractions: Convert contractions like "don't" to "do not"
• Lowercase: Convert text to lowercase
• Remove URLs: Remove web links from text
• Remove newlines: Replace newline characters with spaces
• Remove numbers: Remove digits from text
• Remove punctuation: Remove punctuation marks
• Remove emojis: Remove emoji characters
• Tokenization: Split text into tokens
• Remove stopwords: Remove common words like "the", "a", "is"
• Stemming/Lemmatization: Reduce words to their root forms
• Unicode normalization: Normalize accented characters
• Word length filtering: Filter words by length

Supported Languages

• English

Contributing

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

License

MIT License