FastApi RAG Application

A minimal RAG backend using Python, FastAPI, and vector similarity search. The application supports user authentication, document ingestion, querying, and response feedback.

Features

• User registration and login (JWT-based auth)
• Ask questions and get answers based on ingested documents
• Provide feedback on helpfulness of responses
• Ingest documents for RAG
• Choose between Ollama or Hugging Face embedding models
• Built with FastAPI, PostgreSQL, pgvector, and LangChain

---

Installation

1. Clone the repository

   git clone <your-repo-url>
   cd <repo-folder>

2. Set up a virtual environment

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

3. Install dependencies

   pip install -r requirements.txt

4. Set environment variables Create a .env file in the root directory:

   DATABASE_URL=postgresql+asyncpg://user:password@localhost:5432/yourdb
   JWT_SECRET=your_jwt_secret
   MODEL_FLOW=ollama  # or hf for Hugging Face

5. Run migrations (if using Alembic)

   alembic upgrade head

6. Start the application

   uvicorn app.main:app --reload

---

Embedding Options

You can switch between embedding models by setting the MODEL_FLOW in the .env file:

Ref: .env.sample

Ollama

Uses langchain_ollama:

from langchain_ollama import OllamaEmbeddings

def get_ollama_embeddings(model_name: str = "mxbai-embed-large"):
    return OllamaEmbeddings(model=model_name)

Supported models:

• mxbai-embed-large
• nomic-embed-text

Hugging Face

Uses langchain_huggingface:

from langchain_huggingface import HuggingFaceEmbeddings

def get_hf_embedding_model():
    return HuggingFaceEmbeddings(model_name="sentence-transformers/all-roberta-large-v1")

Supported models:

• sentence-transformers/all-roberta-large-v1 (1024 dims)
• sentence-transformers/all-MiniLM-L6-v2 (384 dims)

---

API Endpoints

Register

curl --location 'http://127.0.0.1:8000/api/v1/auth/register' \
--header 'Content-Type: application/json' \
--data-raw '{
    "email": "test2@example.com",
    "password": "exX$ampd1sfgsdfle",
    "name": "asdf"
}'

Login

curl --location 'http://127.0.0.1:8000/api/v1/auth/login' \
--header 'Content-Type: application/json' \
--data-raw '{
    "email": "test2@example.com",
    "password": "exX$ampd1sfgsdfle"
}'

Ask a Question

curl --location --request GET 'http://localhost:8000/api/v1/ask' \
--header 'accept: application/json' \
--header 'Authorization: Bearer <JWT_TOKEN>' \
--header 'Content-Type: application/json' \
--data '{
    "prompt": "what do you mean by Gen AI?"
}'

Mark Response as Helpful

curl --location --request GET 'http://localhost:8000/api/v1/mark_response' \
--header 'accept: application/json' \
--header 'Authorization: Bearer <JWT_TOKEN>' \
--header 'Content-Type: application/json' \
--data '{
    "is_helpful": true,
    "id": "28cc8a02-7e74-4af9-882f-3a363bd9580e"
}'

Ingest Documents

curl --location 'http://localhost:8000/api/v1/ingest' \
--header 'accept: application/json' \
--header 'Authorization: Bearer <JWT_TOKEN>'

---

Notes

• Replace <JWT_TOKEN> with the token received from the login API.
• You can switch embedding models using MODEL_FLOW = ollama or hf.
• Ensure PostgreSQL and pgvector extension are installed and configured.

---

Tech Stack

• Backend: FastAPI
• Auth: JWT
• Database: PostgreSQL + pgvector
• Embeddings: Ollama / Hugging Face via LangChain
• Vector Store: FAISS / pgvector

---

MIT License

This project is licensed under the MIT License.