api.py

"""
NegaBot API - FastAPI application for tweet sentiment classification
"""
from fastapi import FastAPI, HTTPException
from fastapi.responses import HTMLResponse, Response
from pydantic import BaseModel, Field
from typing import List, Optional
import logging
from datetime import datetime
import json
from model import get_model
from database import log_prediction, get_all_predictions

# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# Initialize FastAPI app
app = FastAPI(
    title="NegaBot API",
    description="NegaBot is a complete sentiment analysis solution that detects positive and negative sentiment in tweets, particularly focusing on product criticism detection. Built with FastAPI, Streamlit, and the powerful jatinmehra/NegaBot-Product-Criticism-Catcher model.",
    version="1.0.0"
)

# Pydantic models for request/response validation
class TweetRequest(BaseModel):
    text: str = Field(..., min_length=1, max_length=1000, description="Tweet text to analyze")
    metadata: Optional[dict] = Field(default=None, description="Optional metadata")

class TweetResponse(BaseModel):
    text: str
    sentiment: str
    confidence: float
    predicted_class: int
    probabilities: dict
    timestamp: str
    request_id: Optional[str] = None

class BatchTweetRequest(BaseModel):
    tweets: List[str] = Field(..., min_items=1, max_items=50, description="List of tweets to analyze")
    metadata: Optional[dict] = Field(default=None, description="Optional metadata")

class BatchTweetResponse(BaseModel):
    results: List[TweetResponse]
    total_processed: int
    timestamp: str

class HealthResponse(BaseModel):
    status: str
    model_loaded: bool
    timestamp: str

# Global variables
model = None

@app.on_event("startup")
async def startup_event():
    """Initialize the model on startup"""
    global model
    try:
        logger.info("Starting NegaBot API...")
        model = get_model()
        logger.info("Model loaded successfully")
    except Exception as e:
        logger.error(f"Failed to load model: {str(e)}")
        raise e

@app.get("/", response_model=dict)
async def root():
    """Root endpoint with API information"""
    return {
        "message": "Welcome to NegaBot API",
        "version": "1.0.0",
        "description": "Tweet Sentiment Classification using NegaBot model",
        "OpenAPI Link": "https://jatinmehra-negabot-api.hf.space/docs",
        "endpoints": {
            "predict": "/predict - Single tweet prediction",
            "batch_predict": "/batch_predict - Multiple tweets prediction",
            "health": "/health - API health check",
            "stats": "/stats - Prediction statistics",
            "dashboard": "/dashboard - Interactive analytics dashboard",
            "dashboard_data": "/dashboard/data - Dashboard data as JSON",
            "download_csv": "/download/predictions.csv - Download predictions as CSV",
            "download_json": "/download/predictions.json - Download predictions as JSON"
        }
    }

@app.get("/health", response_model=HealthResponse)
async def health_check():
    """Health check endpoint"""
    return HealthResponse(
        status="healthy" if model is not None else "unhealthy",
        model_loaded=model is not None,
        timestamp=datetime.now().isoformat()
    )

@app.post("/predict", response_model=TweetResponse)
async def predict_sentiment(request: TweetRequest):
    """
    Predict sentiment for a single tweet
    
    Args:
        request: TweetRequest containing the tweet text
        
    Returns:
        TweetResponse with prediction results
    """
    try:
        if model is None:
            raise HTTPException(status_code=503, detail="Model not loaded")
        
        # Get prediction from model
        result = model.predict(request.text)
        
        # Create response
        response = TweetResponse(
            text=result["text"],
            sentiment=result["sentiment"],
            confidence=result["confidence"],
            predicted_class=result["predicted_class"],
            probabilities=result["probabilities"],
            timestamp=datetime.now().isoformat()
        )
        
        # Log the prediction
        log_prediction(
            text=request.text,
            sentiment=result["sentiment"],
            confidence=result["confidence"],
            metadata=request.metadata
        )
        
        logger.info(f"Prediction made: {result['sentiment']} (confidence: {result['confidence']:.2%})")
        return response
        
    except Exception as e:
        logger.error(f"Error in prediction: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Prediction failed: {str(e)}")

@app.post("/batch_predict", response_model=BatchTweetResponse)
async def batch_predict_sentiment(request: BatchTweetRequest):
    """
    Predict sentiment for multiple tweets
    
    Args:
        request: BatchTweetRequest containing list of tweets
        
    Returns:
        BatchTweetResponse with all prediction results
    """
    try:
        if model is None:
            raise HTTPException(status_code=503, detail="Model not loaded")
        
        # Get predictions for all tweets
        results = model.batch_predict(request.tweets)
        
        # Create response objects
        responses = []
        for result in results:
            response = TweetResponse(
                text=result["text"],
                sentiment=result["sentiment"],
                confidence=result["confidence"],
                predicted_class=result["predicted_class"],
                probabilities=result["probabilities"],
                timestamp=datetime.now().isoformat()
            )
            responses.append(response)
            
            # Log each prediction
            log_prediction(
                text=result["text"],
                sentiment=result["sentiment"],
                confidence=result["confidence"],
                metadata=request.metadata
            )
        
        batch_response = BatchTweetResponse(
            results=responses,
            total_processed=len(responses),
            timestamp=datetime.now().isoformat()
        )
        
        logger.info(f"Batch prediction completed: {len(responses)} tweets processed")
        return batch_response
        
    except Exception as e:
        logger.error(f"Error in batch prediction: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Batch prediction failed: {str(e)}")

@app.get("/stats", response_model=dict)
async def get_prediction_stats():
    """
    Get prediction statistics
    
    Returns:
        Dictionary with prediction statistics
    """
    try:
        predictions = get_all_predictions()
        
        if not predictions:
            return {
                "total_predictions": 0,
                "positive_count": 0,
                "negative_count": 0,
                "average_confidence": 0,
                "message": "No predictions found"
            }
        
        total = len(predictions)
        positive_count = sum(1 for p in predictions if p["sentiment"] == "Positive")
        negative_count = total - positive_count
        avg_confidence = sum(p["confidence"] for p in predictions) / total
        
        stats = {
            "total_predictions": total,
            "positive_count": positive_count,
            "negative_count": negative_count,
            "positive_percentage": round((positive_count / total) * 100, 2),
            "negative_percentage": round((negative_count / total) * 100, 2),
            "average_confidence": round(avg_confidence, 4),
            "last_updated": datetime.now().isoformat()
        }
        
        return stats
        
    except Exception as e:
        logger.error(f"Error getting stats: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Failed to get statistics: {str(e)}")

@app.get("/dashboard/data", response_model=dict)
async def get_dashboard_data():
    """
    Get dashboard data as JSON for API consumption
    """
    try:
        predictions = get_all_predictions()
        
        if not predictions:
            return {
                "metrics": {
                    "total_predictions": 0,
                    "positive_count": 0,
                    "negative_count": 0,
                    "average_confidence": 0
                },
                "recent_predictions": [],
                "message": "No predictions found"
            }
        
        # Calculate metrics
        total = len(predictions)
        positive_count = sum(1 for p in predictions if p["sentiment"] == "Positive")
        negative_count = total - positive_count
        avg_confidence = sum(p["confidence"] for p in predictions) / total
        
        # Get recent predictions (last 20)
        recent_predictions = sorted(predictions, key=lambda x: x["created_at"], reverse=True)[:20]
        
        return {
            "metrics": {
                "total_predictions": total,
                "positive_count": positive_count,
                "negative_count": negative_count,
                "positive_percentage": round((positive_count / total) * 100, 2),
                "negative_percentage": round((negative_count / total) * 100, 2),
                "average_confidence": round(avg_confidence, 4)
            },
            "recent_predictions": recent_predictions,
            "last_updated": datetime.now().isoformat()
        }
        
    except Exception as e:
        logger.error(f"Error getting dashboard data: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Failed to get dashboard data: {str(e)}")

@app.get("/download/predictions.csv")
async def download_predictions_csv():
    """
    Download all predictions as CSV file
    """
    try:
        predictions = get_all_predictions()
        
        if not predictions:
            raise HTTPException(status_code=404, detail="No predictions found to download")
        
        # Convert to pandas DataFrame for easy CSV export
        import pandas as pd
        df = pd.DataFrame(predictions)
        
        # Convert to CSV
        csv_content = df.to_csv(index=False)
        
        # Generate filename with timestamp
        filename = f"negabot_predictions_{datetime.now().strftime('%Y%m%d_%H%M%S')}.csv"
        
        return Response(
            content=csv_content,
            media_type="text/csv",
            headers={"Content-Disposition": f"attachment; filename={filename}"}
        )
        
    except Exception as e:
        logger.error(f"Error downloading CSV: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Failed to download CSV: {str(e)}")

@app.get("/download/predictions.json")
async def download_predictions_json():
    """
    Download all predictions as JSON file
    """
    try:
        predictions = get_all_predictions()
        
        if not predictions:
            raise HTTPException(status_code=404, detail="No predictions found to download")
        
        # Convert to JSON
        json_content = json.dumps(predictions, indent=2, default=str)
        
        # Generate filename with timestamp
        filename = f"negabot_predictions_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json"
        
        return Response(
            content=json_content,
            media_type="application/json",
            headers={"Content-Disposition": f"attachment; filename={filename}"}
        )
        
    except Exception as e:
        logger.error(f"Error downloading JSON: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Failed to download JSON: {str(e)}")

@app.get("/dashboard", response_class=HTMLResponse)
async def dashboard():
    """
    Serve the analytics dashboard as HTML
    """
    try:
        import pandas as pd
        import plotly.express as px
        import plotly.graph_objects as go
        
        # Get prediction data
        predictions = get_all_predictions()
        
        if not predictions:
            html_content = """
            <!DOCTYPE html>
            <html>
            <head>
                <title>NegaBot Dashboard</title>
                <style>
                    body { font-family: Arial, sans-serif; margin: 40px; }
                    .container { max-width: 800px; margin: 0 auto; text-align: center; }
                    .warning { background-color: #fff3cd; border: 1px solid #ffeaa7; padding: 20px; border-radius: 8px; }
                </style>
            </head>
            <body>
                <div class="container">
                    <h1>🤖 NegaBot Analytics Dashboard</h1>
                    <div class="warning">
                        <h3>📭 No prediction data found</h3>
                        <p>Make some predictions using the API first!</p>
                        <p><strong>Quick Start:</strong></p>
                        <ol>
                            <li>Use POST to <code>/predict</code> endpoint</li>
                            <li>Refresh this dashboard to see analytics</li>
                        </ol>
                        <p><strong>Available downloads:</strong></p>
                        <p>
                            <a href="/download/predictions.csv" style="color: #007bff; text-decoration: none;">📥 CSV Format</a> | 
                            <a href="/download/predictions.json" style="color: #007bff; text-decoration: none;">📥 JSON Format</a>
                        </p>
                    </div>
                </div>
            </body>
            </html>
            """
            return HTMLResponse(content=html_content)
        
        # Process data
        df = pd.DataFrame(predictions)
        df['created_at'] = pd.to_datetime(df['created_at'])
        
        # Calculate metrics
        total_predictions = len(df)
        positive_count = len(df[df['sentiment'] == 'Positive'])
        negative_count = total_predictions - positive_count
        avg_confidence = df['confidence'].mean()
        
        # Create sentiment distribution chart
        sentiment_counts = df['sentiment'].value_counts()
        fig_pie = px.pie(
            values=sentiment_counts.values,
            names=sentiment_counts.index,
            title="Sentiment Distribution",
            color_discrete_map={'Positive': '#2E8B57', 'Negative': '#DC143C'}
        )
        pie_html = fig_pie.to_html(include_plotlyjs='cdn', div_id="sentiment-pie")
        
        # Create confidence distribution chart
        fig_hist = px.histogram(
            df,
            x='confidence',
            nbins=20,
            title="Confidence Score Distribution",
            color='sentiment',
            color_discrete_map={'Positive': '#2E8B57', 'Negative': '#DC143C'}
        )
        hist_html = fig_hist.to_html(include_plotlyjs='cdn', div_id="confidence-hist")
        
        # Generate recent predictions table
        recent_df = df.head(10).copy()
        recent_df['text'] = recent_df['text'].str[:100] + '...'
        recent_df['confidence'] = recent_df['confidence'].apply(lambda x: f"{x:.2%}")
        recent_df['created_at'] = recent_df['created_at'].dt.strftime('%Y-%m-%d %H:%M:%S')
        
        table_rows = ""
        for _, row in recent_df.iterrows():
            sentiment_class = "positive" if row['sentiment'] == 'Positive' else "negative"
            table_rows += f"""
            <tr>
                <td>{row['created_at']}</td>
                <td style="max-width: 300px;">{row['text']}</td>
                <td><span class="sentiment {sentiment_class}">{row['sentiment']}</span></td>
                <td>{row['confidence']}</td>
            </tr>
            """
        
        # HTML template
        html_content = f"""
        <!DOCTYPE html>
        <html>
        <head>
            <title>NegaBot Analytics Dashboard</title>
            <meta charset="utf-8">
            <meta name="viewport" content="width=device-width, initial-scale=1">
            <style>
                body {{
                    font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', 'Roboto', sans-serif;
                    margin: 0;
                    padding: 20px;
                    background-color: #f8f9fa;
                }}
                .container {{
                    max-width: 1200px;
                    margin: 0 auto;
                }}
                .header {{
                    text-align: center;
                    color: #1f77b4;
                    margin-bottom: 30px;
                }}
                .metrics-grid {{
                    display: grid;
                    grid-template-columns: repeat(auto-fit, minmax(200px, 1fr));
                    gap: 20px;
                    margin-bottom: 30px;
                }}
                .metric-card {{
                    background: white;
                    padding: 20px;
                    border-radius: 8px;
                    box-shadow: 0 2px 4px rgba(0,0,0,0.1);
                    text-align: center;
                }}
                .metric-value {{
                    font-size: 2em;
                    font-weight: bold;
                    color: #1f77b4;
                }}
                .metric-label {{
                    color: #666;
                    margin-top: 5px;
                }}
                .charts-grid {{
                    display: grid;
                    grid-template-columns: 1fr 1fr;
                    gap: 20px;
                    margin-bottom: 30px;
                }}
                .chart-container {{
                    background: white;
                    padding: 20px;
                    border-radius: 8px;
                    box-shadow: 0 2px 4px rgba(0,0,0,0.1);
                }}
                .table-container {{
                    background: white;
                    padding: 20px;
                    border-radius: 8px;
                    box-shadow: 0 2px 4px rgba(0,0,0,0.1);
                    overflow-x: auto;
                }}
                table {{
                    width: 100%;
                    border-collapse: collapse;
                }}
                th, td {{
                    padding: 12px;
                    text-align: left;
                    border-bottom: 1px solid #eee;
                }}
                th {{
                    background-color: #f8f9fa;
                    font-weight: 600;
                }}
                .sentiment.positive {{
                    background-color: #d4edda;
                    color: #155724;
                    padding: 4px 8px;
                    border-radius: 4px;
                    font-size: 0.9em;
                }}
                .sentiment.negative {{
                    background-color: #f8d7da;
                    color: #721c24;
                    padding: 4px 8px;
                    border-radius: 4px;
                    font-size: 0.9em;
                }}
                .refresh-btn {{
                    background-color: #1f77b4;
                    color: white;
                    border: none;
                    padding: 10px 20px;
                    border-radius: 4px;
                    cursor: pointer;
                    font-size: 14px;
                    margin-bottom: 20px;
                }}
                .refresh-btn:hover {{
                    background-color: #1865a0;
                }}
                .download-btn {{
                    background-color: #28a745;
                    color: white;
                    text-decoration: none;
                    padding: 8px 16px;
                    border-radius: 4px;
                    font-size: 14px;
                    display: inline-block;
                    transition: background-color 0.2s;
                }}
                .download-btn:hover {{
                    background-color: #218838;
                    text-decoration: none;
                    color: white;
                }}
                @media (max-width: 768px) {{
                    .charts-grid {{
                        grid-template-columns: 1fr;
                    }}
                }}
            </style>
        </head>
        <body>
            <div class="container">
                <div class="header">
                    <h1>🤖 NegaBot Analytics Dashboard</h1>
                    <button class="refresh-btn" onclick="location.reload()">🔄 Refresh Data</button>
                </div>
                
                <div class="metrics-grid">
                    <div class="metric-card">
                        <div class="metric-value">{total_predictions}</div>
                        <div class="metric-label">📊 Total Predictions</div>
                    </div>
                    <div class="metric-card">
                        <div class="metric-value">{positive_count}</div>
                        <div class="metric-label">😊 Positive</div>
                    </div>
                    <div class="metric-card">
                        <div class="metric-value">{negative_count}</div>
                        <div class="metric-label">😞 Negative</div>
                    </div>
                    <div class="metric-card">
                        <div class="metric-value">{avg_confidence:.1%}</div>
                        <div class="metric-label">🎯 Avg Confidence</div>
                    </div>
                </div>
                
                <div class="charts-grid">
                    <div class="chart-container">
                        {pie_html}
                    </div>
                    <div class="chart-container">
                        {hist_html}
                    </div>
                </div>
                
                <div class="table-container">
                    <h3>📝 Recent Predictions</h3>
                    <div style="margin-bottom: 15px;">
                        <a href="/download/predictions.csv" class="download-btn" style="margin-right: 10px;">📥 Download CSV</a>
                        <a href="/download/predictions.json" class="download-btn">📥 Download JSON</a>
                    </div>
                    <table>
                        <thead>
                            <tr>
                                <th>Timestamp</th>
                                <th>Tweet Text</th>
                                <th>Sentiment</th>
                                <th>Confidence</th>
                            </tr>
                        </thead>
                        <tbody>
                            {table_rows}
                        </tbody>
                    </table>
                </div>
                
                <div style="text-align: center; margin-top: 30px; color: #666; font-size: 0.9em;">
                    🤖 NegaBot Analytics Dashboard | Last updated: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}
                </div>
            </div>
        </body>
        </html>
        """
        
        return HTMLResponse(content=html_content)
        
    except Exception as e:
        logger.error(f"Error generating dashboard: {str(e)}")
        raise HTTPException(status_code=500, detail=f"Failed to generate dashboard: {str(e)}")

if __name__ == "__main__":
    import uvicorn
    uvicorn.run(app, host="0.0.0.0", port=7860)