DEPLOYMENT.md

# Hugging Face Space Deployment Guide

## 🚀 How to Deploy Your Sentiment Analysis API to Hugging Face Spaces

### Step 1: Create a Hugging Face Account
1. Go to [https://huggingface.co](https://huggingface.co)
2. Sign up for a free account if you don't have one

### Step 2: Create a New Space
1. Go to [https://huggingface.co/new-space](https://huggingface.co/new-space)
2. Fill in the details:
   - **Space name**: `sentiment-analysis-api` (or your preferred name)
   - **License**: MIT
   - **SDK**: Docker
   - **Hardware**: CPU Basic (free tier)

### Step 3: Upload Your Files
Upload all the files from this directory to your new Space:
- `app.py` - FastAPI application
- `sentiment.pkl` - Your trained model
- `requirements.txt` - Python dependencies
- `Dockerfile` - Docker configuration
- `README.md` - Documentation
- `index.html` - Simple web interface (optional)

### Step 4: Your Space Will Auto-Deploy
- Hugging Face will automatically build and deploy your Docker container
- The build process takes 5-10 minutes
- You can monitor the build logs in the Space dashboard

### Step 5: Access Your API
Once deployed, your API will be available at:
- **API Base URL**: `https://YOUR_USERNAME-sentiment-analysis-api.hf.space`
- **Interactive Docs**: `https://YOUR_USERNAME-sentiment-analysis-api.hf.space/docs`
- **Simple Interface**: `https://YOUR_USERNAME-sentiment-analysis-api.hf.space` (if you uploaded index.html)

## 🧪 Testing Locally First

Before deploying, test your API locally:

1. **Install dependencies**:
   ```bash
   pip install -r requirements.txt
   ```

2. **Run the API**:
   ```bash
   python app.py
   ```

3. **Test the endpoints**:
   - Visit `http://localhost:7860/docs` for interactive API docs
   - Run `python test_api.py` to test all endpoints
   - Visit `http://localhost:7860` for the simple web interface

## 📋 API Endpoints

### POST `/predict`
```json
{
  "text": "I love this product!"
}
```
Response:
```json
{
  "prediction": 1,
  "confidence": 0.95,
  "sentiment": "positive"
}
```

### POST `/predict_proba`
```json
{
  "text": "This is terrible"
}
```
Response:
```json
{
  "probabilities": [0.85, 0.15],
  "prediction": 0,
  "sentiment": "negative"
}
```

### POST `/batch_predict`
```json
["I love it!", "This is bad", "It's okay"]
```

## 🔧 Customization

### If your model requires text preprocessing:
Edit the `app.py` file in the prediction functions to add your preprocessing steps:

```python
# Add preprocessing before model.predict()
def preprocess_text(text):
    # Add your preprocessing logic here
    # e.g., tokenization, vectorization, etc.
    return processed_text

# In the prediction functions:
processed_text = preprocess_text(input_data.text)
prediction = model.predict([processed_text])[0]
```

### To change the port or host:
Modify the last line in `app.py`:
```python
uvicorn.run(app, host="0.0.0.0", port=8000)  # Change port as needed
```

## 🆘 Troubleshooting

1. **Model loading errors**: Ensure `sentiment.pkl` is in the same directory
2. **Dependency issues**: Check that all required packages are in `requirements.txt`
3. **Memory issues**: Your model might be too large for the free tier (upgrade to paid tier)
4. **Preprocessing errors**: Make sure your text preprocessing matches your training pipeline

## 💡 Next Steps

- Add authentication for production use
- Implement rate limiting
- Add logging and monitoring
- Create more sophisticated web interface
- Add model versioning