Semantic Search App

A semantic search system that lets you query your documents using natural language. Built with Ollama, FAISS, and FastAPI.

Features

• Document ingestion from .txt files
• Vector-based semantic search using FAISS
• AI-powered responses with context from your documents
• FastAPI web service for easy integration

Prerequisites

1. Install Ollama from ollama.ai
2. Pull required models:

   ollama pull nomic-embed-text
   ollama pull llama3

Installation

1. Install dependencies:

   pip install -r requirements.txt

2. Create docs folder and add your .txt files:

   mkdir docs
   # Add your .txt files to the docs/ folder

Usage

Method 1: FastAPI Web Service

1. Start the server:

   uvicorn main:app --reload

   Server runs at http://localhost:8000

2. Ingest documents:

   curl -X POST "http://localhost:8000/ingest/"

3. Query documents:

   curl -X POST "http://localhost:8000/query/" \
     -H "Content-Type: application/json" \
     -d '{"question": "Your question here"}'

4. Interactive API docs: Visit http://localhost:8000/docs

Method 2: Command Line

1. Ingest documents:

   python ingest.py

2. Query documents:

   python query.py

API Endpoints

POST /ingest/

Processes all .txt files in the docs/ folder and creates vector embeddings.

Response:

{
  "status": "success",
  "message": "5 documents ingested and indexed"
}

POST /query/

Search documents and get AI-generated answers.

Request:

{
  "question": "What are the main topics discussed?"
}

Response:

{
  "question": "What are the main topics discussed?",
  "answer": "Based on the documents, the main topics include..."
}

File Structure

├── requirements.txt    # Dependencies
├── ingest.py          # Document processing
├── query.py           # CLI interface
├── main.py            # FastAPI service
├── docs/              # Your .txt files
├── vector.index       # Generated FAISS index
└── metadata.json      # Generated metadata