Knowledge Base Search Engine

A Retrieval-Augmented Generation (RAG) based search engine that allows users to upload documents (PDF/TXT) and query them for synthesized answers using LLM.

Features

• Document ingestion: Upload multiple PDF and TXT files
• Embeddings: Uses OpenAI embeddings for vectorization
• Vector store: FAISS for efficient similarity search
• LLM: OpenAI GPT for answer synthesis
• API: FastAPI backend
• Frontend: Simple web interface for upload and query

Setup

1. Ensure Python 3.11+ is installed.

2. Clone or download the project.

3. Navigate to the project directory.

4. Create virtual environment:

   py -3 -m venv venv
   

5. Activate the environment:

   call venv\Scripts\activate.bat
   

6. Install dependencies:

   pip install -r requirements.txt
   

7. Set your OpenAI API key:

   set OPENAI_API_KEY=your_api_key_here
   

8. Run the application:

   uvicorn main:app --reload
   

9. Open your browser to http://localhost:5000

Usage

1. Upload documents using the upload form (select multiple PDF/TXT files).
2. Once uploaded, enter a query in the query field and click Search.
3. The synthesized answer will be displayed.

API Endpoints

• POST /upload: Upload files (multipart/form-data)
• POST /query: Query with JSON body {"query": "your question"}
• GET /: Serve the frontend

Evaluation

• Retrieval accuracy: Based on FAISS similarity search
• Synthesis quality: Depends on OpenAI LLM
• Code structure: Modular with separate files for ingestion and retrieval
• LLM integration: Uses LangChain for seamless integration"# Knowledge-base-Search-Engine"