QueryGuard 🔍🛡️

QueryGuard is an end-to-end Search Query Anomaly Detection system built with:

• Isolation Forest for anomaly detection
• FastAPI backend for model inference
• Streamlit frontend for UI
• MLflow (local) for tracking experiments (params, metrics, artifacts)

---

How to Run (Local)

1. Create & activate virtual environment

# Windows
python -m venv myenv
myenv\Scripts\activate

2. Install dependencies

pip install -r requirements-api.txt
pip install -r requirements-frontend.txt
pip install mlflow matplotlib

3. Train model (optional)

python src/train_mlflow.py

4. Start backend (FastAPI)

uvicorn api.main:app --reload
Backend docs:
http://127.0.0.1:8000/docs

5. Start frontend (Streamlit)

streamlit run frontend/app.py
Frontend:
http://127.0.0.1:8501

---

How to Run (DockerHub Images - No Build)

1. Pull images

docker pull shreymishr/queryguard-backend:latest
docker pull shreymishr/queryguard-frontend:latest

2. Run using Docker Compose

#Make sure you have the file docker-compose.prod.yml (present in this repo root).
#From the folder containing it, run:

docker compose -f docker-compose.prod.yml up

3. Open in browser

Frontend:
http://localhost:8501

Backend docs:
http://localhost:8000/docs

#To stop
docker compose -f docker-compose.prod.yml down