FigAGI Engine

The FigAGI (fa) engine integrates multi-document data import functionality and vector database management, combined with the capabilities of the OpenAI model, enhancing the search for private data.

Core Features

Multi-Format Compatibility: Supports the import of various file formats including JSON, TXT, MD, and PDF.
Efficient Vector Retrieval: Provides powerful vector data search functionality.
Smooth Conversational Interaction: Supports continuous conversational interaction experience.

Quick Installation

Install ta as a global command:

pip install -e .

Environment Configuration Steps

Step One: Set Environment Variables

Create .env File: First, copy the .env.example file to a new .env file. This can be done in the terminal using the following command:
```
cp .env.example .env
```
Configure OpenAI Key: Enter your OpenAI key in the .env file.
- The method for obtaining the OpenAI API key can be found in the AGI Classroom Manual.
- Enter your key after OPENAI_API_KEY.
Example:
```
OPENAI_API_KEY='Your OpenAI Key'
```

Step Two: Obtain Weaviate Database Configuration

Register with Weaviate: Visit the Weaviate website, register an account, and log in to create a free vector database.
Configure Database Information: Obtain WEAVIATE_URL and WEAVIATE_API_KEY and fill them in the .env file.

Note: The free service is valid for 14 days.

Step Three: Import Data

Use the following command to import data into the TA engine:

ta import --path data

Step Four: Start Web Service

Start the local web service with the following command:

python server/web.py

The service will run on the local port http://127.0.0.1:7860.

Additional Information

Configure the BYPASS_AUTH=1 environment variable to run the program without using LDAP authentication.

wsdo / figagi