Smart-X FAQ Bot

A FastAPI + Streamlit application for classifying user queries into intents using ML models like SVM and CatBoost. It supports training on custom data, intent prediction, and interactive querying via a web interface.

---

📁 Project Structure

smart-x-faq-bot/
├── app/
│   ├── controller/
│   │   └── intent_controller.py       # API endpoints for predict, train, and model list
│   ├── core/
│   │   └── config.py                  # Constants and config loading
│   ├── service/
│   │   ├── intent_service.py         # Logic to run predictions
│   │   └── training.py               # Model training logic
│   └── __init__.py
├── architecture/
│   ├── architecture.md               # Architecture description and explanation
│   └── smart-x-faq-bot.svg           # SVG diagram of system architecture
├── data/
│   ├── uploaded_data/                # Uploaded training files
│   ├── intent_answers.json           # Mapping of intent to answer
|   └── sample_sports_data.xlsx          # Sample training data
├── frontend/
│   └── main.py                       # Streamlit UI for training and prediction
├── models/                           # Auto-generated after training
├── main.py                           # FastAPI app entry
├── run.py                            # Runs both FastAPI and Streamlit apps
├── requirements.txt                  # Python dependencies
├── Dockerfile                        # Containerization setup
└── .gitignore

📁 Architecture

• Architecture: Read detailed architecture explanation here

---

🚀 Features

• Upload training data (.xlsx) with Question and Label columns.
• Choose and train models (SVM or CatBoost).
• View available trained models.
• Predict intent from user queries.
• Streamlit UI for ease of interaction.
• REST API support for integration.

---

📦 Installation

1. Clone and Set Up

git clone <repo-url>
cd <repo-folder>
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

2. Run Locally

python run.py

• FastAPI runs at: http://localhost:8001/docs
• Streamlit UI at: http://localhost:8501

---

📊 Sample Data Format

Upload a .xlsx file with the following structure:

Question	Label
How many players in a team?	team_info
What is the duration?	match_rules

Use sample_sports_data.xlsx as a reference.

---

🔧 API Endpoints

Method	Endpoint	Description
POST	/api/predict	Predicts intent from a query
GET	/api/models	Lists all available models
POST	/api/train	Trains model with uploaded file

---

🐳 Docker

# Build
docker build -t smart-x-faq-bot .

# Run with volume mounts
docker run -p 8001:8001 -p 8501:8501 \
  -v $(pwd)/models:/app/models \
  -v $(pwd)/data/uploaded_data:/app/data/uploaded_data \
  smart-x-faq-bot

---

🧠 Model Logic

• Uses the pretrained transformer model all-mpnet-base-v2 from sentence-transformers to embed queries into semantic vector space.

• Classifies intents using:

  • SVM: GridSearchCV-tuned Support Vector Classifier (linear kernel)
  • CatBoost: Optimized gradient boosting classifier for multiclass prediction

• Saves models to /models/<model_type>_<timestamp>/

---