ContextoModel

ContextoModel is an advanced, context-aware language model with function-calling capabilities, conversation history tracking, and stream-based response generation. It uses n-gram based prediction to generate contextually appropriate responses and can be trained on various contexts.

Features

• Context-aware language modeling
• N-gram based prediction with variable length
• Conversation history tracking
• Stream-based response generation
• Function calling integration
• Debug mode for detailed logging
• Easy to train and use
• Save and load model states

Installation

1. Clone this repository or copy the ContextoModel.js file into your project.
2. Ensure you have Node.js installed on your system.
3. Install the required dependencies:

npm install

Usage

Importing the Model

const ContextoModel = require('./path/to/ContextoModel');

Creating a New Model

const model = new ContextoModel(3, ['casual', 'formal'], 5);

This creates a new model with trigrams (n=3), two contexts: 'casual' and 'formal', and a history size of 5.

Training the Model

model.train("This is a sample sentence.", 'casual');

Generating Responses

const response = model.generateResponse('casual', "This is a", 20);
console.log(response);

Using Stream-based Response Generation

const responseGenerator = model.generateResponseStream('casual', "This is a", 20);
for (const partialResponse of responseGenerator) {
    console.log(partialResponse);
}

Registering Functions

model.registerFunction('getTime', () => new Date().toLocaleTimeString(), 'Get current time');

Saving and Loading the Model

// Saving
model.save('model.json');

// Loading
model.load('model.json');

Enabling Debug Mode

model.debugMode = true;

API Reference

constructor(n, contexts = ['default'], historySize = 5)

Creates a new ContextoModel instance.

• n: The n-gram size
• contexts: An array of context names
• historySize: Number of previous interactions to consider

train(text, context = 'default')

Trains the model on the given text in the specified context.

predict(context, phrase, excludeWords = [])

Predicts the next word given a context and a phrase.

registerFunction(name, func, description)

Registers a function that can be called during response generation.

generateResponse(context, input, maxLength = 20)

Generates a complete response given a context and an input phrase.

*generateResponseStream(context, input, maxLength = 20)

Generates a response stream, yielding each part of the response as it's generated.

filterResponse(response, input)

Filters the generated response to avoid repetition and improve coherence.

save(filename)

Saves the current model state to a file.

load(filename)

Loads a model state from a file.

License

MIT License

Contributing

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

Support

If you have any questions or need help with using ContextoModel, please open an issue in the GitHub repository.