AGIMETTA

AGIMETTA is a project that integrates two AI techniques within a unified framework using the Hyperon framework. The project demonstrates how Python functions are used with MeTTa to perform specific tasks.

The Agents

News Agent

• Purpose: Fetches news articles from NewsAPI.
• Functionality: In AGIMETTA, the News Agent fetches a news article based on a query parameter from NewsAPI and processes the description for further use.

Sentiment Analysis Agent

• Purpose: Provides sentiment analysis based on the provided text description.
• Functionality: This agent analyzes the sentiment of the input text and returns whether the sentiment is positive, negative, or neutral.

MeTTa Agent

• Purpose: Acts as the symbolic reasoner agent.
• Functionality: Calls the News and Sentiment agents, stores and reads their outputs, and performs symbolic reasoning based on the results.

Motivation

The AGIMETTA project aims to demonstrate how the Hyperon framework can be used as a unified environment to perform neural-symbolic reasoning. It shows how different Python functions can be defined and utilized as grounded atoms within MeTTa, the programming language of the Hyperon framework.

Setup Guide

1. Clone the repository:

   git clone https://github.com/kecheste/agimetta.git

2. Navigate to the repository:

   cd agimetta

3. Install the requirements:

   pip install -r requirements.txt

4. Run the MeTTa file:

   metta run-agimetta.metta

Example Usage

To use the News and Sentiment agents in MeTTa, you can run the following commands in the MeTTa file:

!(import! &self agimetta)
!(news "Tesla")
!(sentiment "Tesla is revolutionizing the electric vehicle market with innovative technology.")
! (add-reduct &self (news "Tesla"))
!(match &self $x $x)