#!/usr/bin/env python3
"""
RF-DETR SoccerNet - Complete Usage Examples

This script demonstrates how to use the RF-DETR SoccerNet model for soccer analysis.
Shows image processing, data analysis, and visualization using included sample files.
"""

import os
import sys
import pandas as pd
import numpy as np
from pathlib import Path

# Import our RF-DETR SoccerNet class
from inference import RFDETRSoccerNet


def basic_image_analysis():
    """Basic image processing example using included sample."""
    print("🖼️  BASIC IMAGE ANALYSIS EXAMPLE")
    print("=" * 50)
    
    # Initialize model (automatically detects CUDA/CPU)
    print("Loading RF-DETR SoccerNet model...")
    model = RFDETRSoccerNet()
    
    # Process sample image (included in repository)
    image_path = "examples/sample_soccer_frame.jpg"
    
    if not os.path.exists(image_path):
        print(f"⚠️  Sample image not found: {image_path}")
        print("Please run this script from the repository root directory.")
        return None
    
    print(f"Processing sample image: {image_path}")
    
    # Run inference on single image
    df = model.process_image(
        image_path=image_path,
        confidence_threshold=0.5
    )
    
    # Display results
    print(f"\n📊 ANALYSIS RESULTS")
    print(f"Total detections: {len(df):,}")
    
    if len(df) > 0:
        # Show detections by class
        print(f"\n🎯 Detections by class:")
        class_stats = df['class_name'].value_counts()
        for class_name, count in class_stats.items():
            avg_conf = df[df['class_name'] == class_name]['confidence'].mean()
            print(f"  {class_name}: {count} (avg confidence: {avg_conf:.3f})")
        
        # Save results
        print(f"\n💾 Saving results...")
        model.save_results(df, "image_analysis.csv", format="csv")
        model.save_results(df, "image_analysis.json", format="json")
    else:
        print("No detections found. Try lowering the confidence threshold.")
    
    return df


def analyze_sample_data():
    """Analyze sample detection data included in repository."""
    print("\n📊 SAMPLE DATA ANALYSIS EXAMPLE")
    print("=" * 50)
    
    # Load sample detection data
    csv_path = "examples/sample_detections.csv"
    
    if not os.path.exists(csv_path):
        print(f"⚠️  Sample data not found: {csv_path}")
        print("Please run this script from the repository root directory.")
        return None
    
    print(f"Loading sample detection data: {csv_path}")
    df = pd.read_csv(csv_path)
    
    # Basic statistics
    print(f"\n📈 Dataset Statistics:")
    print(f"  Total detections: {len(df):,}")
    print(f"  Frames: {df['frame'].nunique()}")
    print(f"  Time span: {df['timestamp'].min():.2f}s - {df['timestamp'].max():.2f}s")
    
    # Class distribution
    print(f"\n🎯 Class Distribution:")
    class_counts = df['class_name'].value_counts()
    for class_name, count in class_counts.items():
        percentage = (count / len(df)) * 100
        avg_conf = df[df['class_name'] == class_name]['confidence'].mean()
        print(f"  {class_name}: {count} ({percentage:.1f}%) | Avg confidence: {avg_conf:.3f}")
    
    # Confidence analysis
    print(f"\n📊 Confidence Analysis:")
    print(f"  Overall mean: {df['confidence'].mean():.3f}")
    print(f"  Overall std: {df['confidence'].std():.3f}")
    print(f"  Range: {df['confidence'].min():.3f} - {df['confidence'].max():.3f}")
    
    # Object size analysis
    print(f"\n📏 Object Size Analysis:")
    df['area'] = df['width'] * df['height']
    for class_name in df['class_name'].unique():
        class_data = df[df['class_name'] == class_name]
        avg_area = class_data['area'].mean()
        print(f"  {class_name}: {avg_area:.1f} px² average area")
    
    return df


def ball_possession_example():
    """Demonstrate ball possession analysis using sample data."""
    print("\n⚽ BALL POSSESSION ANALYSIS EXAMPLE") 
    print("=" * 50)
    
    # Load sample data
    csv_path = "examples/sample_detections.csv"
    if not os.path.exists(csv_path):
        print(f"⚠️  Sample data not found: {csv_path}")
        return None
    
    df = pd.read_csv(csv_path)
    
    # Initialize model for possession analysis
    model = RFDETRSoccerNet()
    
    print("Analyzing ball possession...")
    possession_df = model.analyze_ball_possession(df, distance_threshold=150)
    
    if len(possession_df) > 0:
        print(f"\n📈 Possession Analysis Results:")
        print(f"  Possession events: {len(possession_df)}")
        print(f"  Average distance to ball: {possession_df['distance_to_ball'].mean():.1f} pixels")
        print(f"  Frames with possession: {possession_df['frame'].nunique()}")
        
        # Save possession analysis
        model.save_results(possession_df, "ball_possession_analysis.csv")
        print(f"💾 Possession analysis saved to: ball_possession_analysis.csv")
    else:
        print("No possession events found. Try increasing the distance threshold.")
    
    return possession_df


def visualization_example():
    """Show how to work with detection visualizations."""
    print("\n🎨 VISUALIZATION EXAMPLE")
    print("=" * 50)
    
    # Check if visualization example exists
    viz_path = "examples/detection_visualization.jpg"
    
    if os.path.exists(viz_path):
        print(f"✅ Sample visualization available: {viz_path}")
        print("This image shows detected objects with bounding boxes and labels.")
        print("Colors: Red=Ball, Green=Player, Yellow=Referee, Cyan=Goalkeeper")
        
        # Get file size
        file_size = os.path.getsize(viz_path) / (1024 * 1024)  # MB
        print(f"📊 Visualization size: {file_size:.2f} MB")
    else:
        print(f"⚠️  Visualization not found: {viz_path}")
        print("You can create visualizations by drawing bounding boxes on images.")
    
    # Example of how to create visualizations programmatically
    print(f"\n💡 To create your own visualizations:")
    print("1. Use cv2.rectangle() to draw bounding boxes")  
    print("2. Use cv2.putText() to add labels")
    print("3. Use different colors for each class")
    print("4. Save with cv2.imwrite()")


def advanced_usage_tips():
    """Show advanced usage patterns and tips."""
    print("\n🚀 ADVANCED USAGE TIPS")
    print("=" * 50)
    
    print("📚 Model Configuration:")
    print("  - Use confidence_threshold=0.3 for more detections")
    print("  - Use confidence_threshold=0.7 for higher precision")
    print("  - Process every N frames: frame_skip=N for faster analysis")
    
    print(f"\n⚡ Performance Tips:")
    print("  - GPU acceleration: Model automatically uses CUDA if available")
    print("  - Batch processing: Process multiple images in sequence")
    print("  - Memory: Large videos may need max_frames limit")
    
    print(f"\n📊 Data Export Options:")
    print("  - CSV: Best for spreadsheet analysis")
    print("  - JSON: Best for web applications")
    print("  - Parquet: Best for large datasets")
    
    print(f"\n🎯 Use Cases:")
    print("  - Player tracking and formation analysis")
    print("  - Ball possession and movement statistics") 
    print("  - Referee positioning analysis")
    print("  - Automated highlight generation")
    print("  - Broadcast enhancement")


def main():
    """Run all examples in sequence."""
    print("🏆 RF-DETR SOCCERNET - COMPLETE EXAMPLES")
    print("=" * 60)
    print("This script demonstrates all features using included sample files.")
    print("No external videos or images required!")
    print("=" * 60)
    
    try:
        # Run all examples
        image_results = basic_image_analysis()
        sample_analysis = analyze_sample_data()
        possession_results = ball_possession_example()
        visualization_example()
        advanced_usage_tips()
        
        print(f"\n🎉 ALL EXAMPLES COMPLETED SUCCESSFULLY!")
        print("Check the generated files:")
        print("  - image_analysis.csv")
        print("  - image_analysis.json") 
        print("  - ball_possession_analysis.csv")
        
    except Exception as e:
        print(f"\n❌ Error running examples: {e}")
        print("Make sure you're running from the repository root directory.")
        import traceback
        traceback.print_exc()


if __name__ == "__main__":
    main()