FastAPI DocGPT: FastAPI-based PDF document processing and Q&A system

General Introduction

FastAPI DocGPT is a FastAPI-based document quiz system that allows users to upload PDF files and quiz them based on the document content. The system uses OpenAI's embedding technology to embed document content into a vector database, Qdrant, to realize the intelligent Q&A function. Users can upload documents and ask questions through the API interface, and the system will return intelligent answers based on the document content.

Function List

• PDF Upload: Users can upload PDF files, which are processed and stored in the vector database.
• question and answer system: Users can ask questions based on the uploaded PDF content and the system will return intelligent answers.
• API Documentation: Provides auto-generated API documentation through Swagger for developers' convenience.
• Cross-domain resource sharing: Supports CORS, allowing front-end requests from different domains.

Using Help

Installation and Setup

1. clone warehouse

   git clone https://github.com/shaheryaryousaf/fastapi-docgpt
   cd fastapi-docgpt
   

2. Setting up a virtual environment

   python3 -m venv venv
   source venv/bin/activate # Windows: venv\Scripts\activate
   

3. Installation of dependencies

   pip install -r requirements.txt
   

4. Configuring Environment Variables In the project root directory, create the .env file and add API keys for OpenAI and Qdrant:

   OPENAI_API_KEY=your-openai-api-key
   QDRANT_URL=your-qdrant-url
   QDRANT_API_KEY=your-qdrant-api-key
   

Guidelines for use

1. Launching the FastAPI application

   uvicorn app:app --reload
   

   This will start the FastAPI application with the http://127.0.0.1:8000 Provision of services.

2. Upload PDF files pass (a bill or inspection etc) /upload-pdf/ Endpoints upload PDF files:

   curl -X POST "http://127.0.0.1:8000/upload-pdf/" -F "file=@yourfile.pdf"
   

   The system processes the PDF file and embeds its contents into the vector database.

3. put pass (a bill or inspection etc) /ask-question/ Endpoints raise questions:

   curl -X POST "http://127.0.0.1:8000/ask-question/" -H "Content-Type: application/json" -d '{"question": "your question"}'
   

   The system returns intelligent answers based on the content of the document.

Project structure

• app.py: The main FastAPI application file containing the API endpoints for the PDF upload and Q&A system.
• utils.py: Contains utility functions for processing PDF files, sending embeddings to vector databases, and retrieving answers from embeddings.
• .env File: Manages API keys for OpenAI and Qdrant.

dependency library (computing)

• FastAPI: Used to build Web APIs.
• Qdrant Client: Used to store and retrieve document embeds.
• LangChain: For processing PDF and embedding.
• OpenAI: for generating embeddings and AI model responses.
• PyPDFLoader: Used to extract text from PDF files.
• CORS Middleware: Handles cross-domain resource sharing (CORS), allowing front-end requests from different domains.
• dotenv: Manage environment variables (such as API keys).