chat-agent

Overview

The Chat-Agent system is designed to execute GPT queries using intelligent tool selection. It leverages a hierarchical approach, dynamically choosing between summarization and Retrieval-Augmented Generation (RAG) tools, further refined to specific implementations within each category.

System Components

Function Maps: Two primary mappings, SUMMARIZATION_FUNC_MAP and RAG_FUNC_MAP, direct the system to the appropriate processing function based on query characteristics.
Tool Builder: Constructs tool objects for the completions API, enabling dynamic function selection through GPT-4-Turbo completions.
Tool Selection: Utilizes OpenAI's API to select the optimal tool for query execution, based on the system-generated context and the user's query.

Operation

Initialization: The system categorizes the query as either requiring summarization or RAG, based on predefined logic and tool mappings.
Tool Selection: Through the select_tool function, it determines the specific tool (e.g., rag_metagpt, rag_autogen, etc.) that best suits the query.
Execution: The selected tool processes the query, with the outcome logged for transparency and debugging purposes.

Development

setup environment and run demo

virtualenv venv

source venv/bin/activate

pip install -r requirements.txt

export OPENAI_API_KEY=your-api-key

python src/demo.py

Run tests

pip install -r requirements.dev.txt

pytest

voynow / chat-agent