https://github.com/balaji-r-05/askdocs-ai
An AI-powered chatbot that leverages RAG (Retrieval-Augmented Generation) to answer your questions based on the content of uploaded PDFs
https://github.com/balaji-r-05/askdocs-ai
chromadb groq-cloud langchain python streamlit
Last synced: about 2 months ago
JSON representation
An AI-powered chatbot that leverages RAG (Retrieval-Augmented Generation) to answer your questions based on the content of uploaded PDFs
- Host: GitHub
- URL: https://github.com/balaji-r-05/askdocs-ai
- Owner: Balaji-R-05
- Created: 2025-06-17T17:17:37.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2025-06-17T17:55:35.000Z (about 1 year ago)
- Last Synced: 2025-06-17T18:46:32.922Z (about 1 year ago)
- Topics: chromadb, groq-cloud, langchain, python, streamlit
- Language: Python
- Homepage:
- Size: 0 Bytes
- Stars: 1
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# AskDocs AI: AI-Powered PDF Q&A Bot
**AskDocs AI** is an AI-powered chatbot that leverages **Hybrid RAG (Retrieval-Augmented Generation)** to answer your questions based on the content of uploaded PDFs. It combines semantic vector search with traditional keyword-based search for superior accuracy.
## Key Features
- **Hybrid Search**: Combines **ChromaDB** (semantic) and **BM25** (keyword) retrieval.
- **LLM Powered**: High-performance LLM via Groq Cloud.
- **Async Processing**: PDF ingestion and indexing are offloaded to background threads.
- **Multimodal Support**: Optimized for PDF extraction and processing.
## Tech Stack
- **Backend:** FastAPI, LangChain (Classic), ChromaDB, Groq Cloud
- **Frontend:** Streamlit
- **Search Engines:** BM25 (Keyword), Vector (Cosine Similarity)
- **Embeddings:** HuggingFace (Sentence Transformers)
- **Containerization:** Docker & Docker Compose
## Configuration
Control the behavior of the Hybrid Search by adjusting weights in your `.env` file or `server/config.py`:
| Variable | Description | Default |
|----------|-------------|---------|
| `HYBRID_SEARCH_BM25_WEIGHT` | Weight for keyword search (0.0 to 1.0) | `0.5` |
| `HYBRID_SEARCH_CHROMA_WEIGHT` | Weight for semantic search (0.0 to 1.0) | `0.5` |
| `GROQ_API_KEY` | Your Groq Cloud API Key | *Required* |
## Optimization Features
- **Split Dependencies**: Client and Server have separate requirement files to minimize image sizes.
- **CPU-Only Optimization**: Server image is optimized for CPU-only environments, reducing size from ~12.8GB to ~2.3GB.
- **Persistent Memory**: Uses Docker volumes to persist the ChromaDB vector store and uploaded files.
## Setup Instruction
### 1. Set up environment variables
Create a `.env` file in the root directory:
```env
GROQ_API_KEY=your_api_key_here
```
### 2. Run with Docker (Recommended)
```sh
docker-compose up -d --build
```
- **Streamlit UI**: [http://localhost:8501](http://localhost:8501)
- **FastAPI Docs**: [http://localhost:8000/docs](http://localhost:8000/docs)
---
### Alternative: Local Setup
1. **Create and activate a virtual environment**
```sh
python -m venv venv
.\venv\Scripts\activate # Windows
```
2. **Install dependencies**
```sh
pip install -r requirements.client.txt
pip install -r requirements.server.txt
```
3. **Run the services**
```sh
# Backend
python server/main.py
# Frontend
streamlit run client/main.py
```
## Testing & Verification
To verify that the Hybrid Search mechanism and LLM integration are working correctly:
```sh
python server/tests/test_hybrid_search.py
```
This script validates:
1. Vectorstore connectivity.
2. BM25 index reconstruction.
3. Ensemble Retriever initialization.