Ottopia AI Agent

A flexible AI research assistant built from scratch using Python and LangChain.

Overview

This project implements an AI agent with memory capabilities, structured reasoning, and various tools for research tasks. The agent can:

• Answer queries with structured responses
• Maintain conversation memory
• Search the web using DuckDuckGo
• Query Wikipedia for information
• Read PDF documents
• Execute shell commands
• Save research outputs to text files

Setup

1. Clone the repository
2. Install dependencies using Make:

   make install
   

   This will create a virtual environment and install all required packages.
3. Create a .env file based on sample.env with your API keys

Usage

Run the assistant using Make:

make run

Available Commands

• help: Show help message
• exit, quit, bye: End the conversation
• memory: Show memory summary
• clear memory: Clear agent's memory (with backup)
• friendly mode: Toggle character-by-character output

API Server

The project includes a REST API server that allows you to interact with the AI agent programmatically or through a web interface.

Starting the API Server

make run-api

This will start the FastAPI server on http://localhost:8000. The API includes automatic interactive documentation accessible at http://localhost:8000/docs.

API Endpoints

• POST /session: Create a new session with the agent
• POST /query: Send a query to the agent
• POST /command: Execute a command on the agent
• GET /health: Check the health of the API server

Using the Command-Line Client

A simple command-line client is provided for testing the API:

make run-client

Testing the API

Run the automated tests with pytest:

make test-api

This runs a suite of unit tests that verify the API's functionality, including session creation, query processing, and command execution.

Project Structure

• src/agent/: Agent implementation
• src/memory/: Memory management
• src/reasoning/: Structured reasoning components
• src/tools/: Tool implementations (search, Wikipedia, shell, etc.)
• main.py: Entry point for the CLI application
• api.py: FastAPI server for the agent
• client.py: Command-line client for interacting with the API
• test_api.py: Test suite for the API server

Dependencies

• LangChain
• Wikipedia API
• DuckDuckGo Search
• OpenAI/Anthropic integration
• Python-dotenv
• Pydantic
• FastAPI

License

[MIT License]