import gradio as gr
import pandas as pd
import plotly.graph_objects as go
import plotly.express as px
from plotly.subplots import make_subplots
import numpy as np
import json
import os
from typing import Dict, List, Any, Tuple, Optional
import tempfile
import uuid
import warnings
warnings.filterwarnings('ignore')

# Initialize Claude API
def get_claude_client():
    """Initialize Claude API client without proxies"""
    api_key = os.getenv("CLAUDE_API_KEY")
    if not api_key:
        print("CLAUDE_API_KEY not set. Falling back to default recommendations.")
        return None
    try:
        import anthropic
        return anthropic.Anthropic(api_key=api_key)
    except Exception as e:
        print(f"Error initializing Claude client: {e}")
        return None

def convert_numpy_types(obj: Any) -> Any:
    """Recursively convert NumPy types to Python native types"""
    if isinstance(obj, np.integer):
        return int(obj)
    elif isinstance(obj, np.floating):
        return float(obj)
    elif isinstance(obj, np.ndarray):
        return obj.tolist()
    elif isinstance(obj, dict):
        return {key: convert_numpy_types(value) for key, value in obj.items()}
    elif isinstance(obj, list):
        return [convert_numpy_types(item) for item in obj]
    elif isinstance(obj, tuple):
        return tuple(convert_numpy_types(item) for item in obj)
    return obj

class DataProcessor:
    """Handles data loading and processing operations"""
    
    @staticmethod
    def load_and_clean_data(file_path: str) -> Optional[pd.DataFrame]:
        """Load and clean CSV/Excel data"""
        try:
            if not os.path.exists(file_path):
                raise FileNotFoundError(f"File {file_path} not found")
                
            if file_path.endswith('.csv'):
                df = pd.read_csv(file_path)
            else:
                df = pd.read_excel(file_path)
            
            # Clean column names
            df.columns = df.columns.str.strip()
            
            # Ensure required columns exist
            required_columns = ['CreativeID', 'Title', 'ReportingBrand']
            missing_columns = [col for col in required_columns if col not in df.columns]
            
            if missing_columns:
                print(f"Missing required columns: {missing_columns}")
                return None
            
            return DataProcessor.calculate_creativity_scores(df)
            
        except Exception as e:
            print(f"Error loading file: {e}")
            return None
    
    @staticmethod
    def calculate_creativity_scores(df: pd.DataFrame) -> pd.DataFrame:
        """Calculate comprehensive creativity scores with optimized operations"""
        df = df.copy()
        df['creativity_score'] = 0
        
        # Cache emotion columns
        emotion_cols = [col for col in df.columns if any(emotion in col for emotion in ['Happiness', 'Surprise', 'Sadness', 'Fear', 'Anger'])]
        if emotion_cols:
            df['emotional_range'] = df[emotion_cols].std(axis=1, skipna=True)
            df['creativity_score'] += df['emotional_range'].fillna(0) * 3
        
        # Performance metrics with vectorized operations
        score_mappings = {
            'EIA': 8,
            'DecimalStar': 6,
            'FinalFluency': lambda x: (x.fillna(0) / 15) * 4,
            'Spike': 5,
            'TotalDynamism': lambda x: (x.fillna(0) / 10) * 3,
            'Duration': lambda x: np.log1p(x.fillna(30)) * 2
        }
        
        for col, weight in score_mappings.items():
            if col in df.columns:
                if callable(weight):
                    df['creativity_score'] += weight(df[col])
                else:
                    df['creativity_score'] += df[col].fillna(0) * weight
        
        return df
    
    @staticmethod
    def get_creative_ids_list(df: pd.DataFrame) -> List[str]:
        """Get list of Creative IDs for dropdown"""
        if df is None or 'CreativeID' not in df.columns:
            return []
        return sorted(df['CreativeID'].unique().tolist())
    
    @staticmethod
    def get_campaign_data(df: pd.DataFrame, creative_id: str) -> Optional[Dict]:
        """Get comprehensive campaign data for analysis"""
        if df is None or creative_id not in df['CreativeID'].values:
            return None
        
        campaign = df[df['CreativeID'] == creative_id].iloc[0]
        
        analysis_data = {
            'creative_id': campaign.get('CreativeID', 'N/A'),
            'title': campaign.get('Title', 'Untitled Campaign'),
            'brand': campaign.get('ReportingBrand', 'Unknown Brand'),
            'advertiser': campaign.get('ReportingAdvertiser', campaign.get('ReportingBrand', 'Unknown')),
            'category': campaign.get('ReportingSubCategory', 'General'),
            'market': campaign.get('Market', 'Unknown'),
            'duration': campaign.get('Duration', 30),
            'creativity_score': campaign.get('creativity_score', 0),
            'performance_metrics': {
                'eia': campaign.get('EIA', 0),
                'decimal_star': campaign.get('DecimalStar', 0),
                'final_fluency': campaign.get('FinalFluency', 0),
                'fast_fluency': campaign.get('FastFluency', 0),
                'spike': campaign.get('Spike', 0),
                'oei': campaign.get('OEI', 0),
                'total_dynamism': campaign.get('TotalDynamism', 0),
                'dynamism_index': campaign.get('DynamismIndex', 0)
            },
            'emotional_profile': {
                'contempt': campaign.get('Contempt (%)', 0),
                'disgust': campaign.get('Disgust (%)', 0),
                'anger': campaign.get('Anger (%)', 0),
                'fear': campaign.get('Fear (%)', 0),
                'sadness': campaign.get('Sadness (%)', 0),
                'neutral': campaign.get('Neutral (%)', 0),
                'happiness': campaign.get('Happiness (%)', 0),
                'surprise': campaign.get('Surprise (%)', 0)
            },
            'sample_info': {
                'count_completes': campaign.get('CountCompletes', 0),
                'sample_description': campaign.get('SampleDescription', 'General Population'),
                'sample_details': campaign.get('SampleDetails', 'Mixed Demographics')
            },
            'technical_details': {
                'display_type': campaign.get('DisplayType', 'Standard'),
                'survey_type': campaign.get('SurveyType', 'Online'),
                'stimuli_type': campaign.get('StimuliType', 'Video'),
                'advert_medium': campaign.get('AdvertMedium', 'Digital'),
                'advert_format': campaign.get('AdvertFormat', '30s Spot')
            }
        }
        
        return convert_numpy_types(analysis_data)

class ChartGenerator:
    """Generates professional charts with dark pink theme"""
    
    @staticmethod
    def create_executive_dashboard(data: Dict) -> go.Figure:
        """Create comprehensive executive dashboard with fixed text overlap"""
        fig = make_subplots(
            rows=3, cols=4,
            subplot_titles=[''] * 12,
            specs=[[{'type': 'indicator'}] * 4] * 3,
            vertical_spacing=0.15,
            horizontal_spacing=0.1
        )
        
        # Dark pink color scheme
        primary_color = "#C2185B"
        secondary_color = "#E91E63"
        accent_color = "#F8BBD9"
        
        # Custom titles
        titles = [
            '🎯 Creativity Score', '⚡ EIA Performance', '🎭 Fluency Rating', '📈 Spike Factor',
            '💫 Emotional Impact', '🔄 Dynamism Index', '⭐ Star Rating', '⏱️ Duration',
            '👥 Sample Size', '🌍 Market Score', '🏢 Brand Power', '📊 Category Score'
        ]
        
        # Row 1 - Core Performance
        fig.add_trace(go.Indicator(
            mode="gauge+number+delta",
            value=round(data['creativity_score'], 1),
            domain={'x': [0, 1], 'y': [0, 1]},
            title={'text': titles[0], 'font': {'size': 12, 'color': primary_color, 'family': 'Arial'}},
            gauge={
                'axis': {'range': [None, 100], 'tickcolor': primary_color},
                'bar': {'color': primary_color, 'thickness': 0.8},
                'bgcolor': "rgba(248, 187, 217, 0.2)",
                'borderwidth': 2,
                'bordercolor': accent_color,
                'steps': [
                    {'range': [0, 30], 'color': '#FCE4EC'},
                    {'range': [30, 60], 'color': '#F8BBD9'},
                    {'range': [60, 80], 'color': '#F48FB1'},
                    {'range': [80, 100], 'color': accent_color}
                ],
                'threshold': {
                    'line': {'color': secondary_color, 'width': 4},
                    'thickness': 0.8,
                    'value': 75
                }
            },
            number={'font': {'size': 20, 'color': primary_color}},
            delta={'reference': 50, 'relative': True, 'font': {'size': 14}}
        ), row=1, col=1)
        
        fig.add_trace(go.Indicator(
            mode="gauge+number",
            value=round(data['performance_metrics']['eia'], 2),
            title={'text': titles[1], 'font': {'size': 12, 'color': primary_color, 'family': 'Arial'}},
            gauge={
                'axis': {'range': [0, 10], 'tickcolor': primary_color},
                'bar': {'color': secondary_color, 'thickness': 0.7},
                'bgcolor': "rgba(194, 24, 91, 0.1)",
                'steps': [
                    {'range': [0, 3], 'color': '#FCE4EC'},
                    {'range': [3, 6], 'color': '#F8BBD9'},
                    {'range': [6, 8], 'color': '#F48FB1'},
                    {'range': [8, 10], 'color': accent_color}
                ]
            },
            number={'font': {'size': 18, 'color': secondary_color}}
        ), row=1, col=2)
        
        fig.add_trace(go.Indicator(
            mode="number+delta",
            value=round(data['performance_metrics']['final_fluency'], 1),
            title={'text': titles[2], 'font': {'size': 12, 'color': primary_color, 'family': 'Arial'}},
            delta={'reference': 60.0, 'relative': True, 'font': {'color': secondary_color}},
            number={'font': {'size': 18, 'color': primary_color}, 'suffix': '%'}
        ), row=1, col=3)
        
        fig.add_trace(go.Indicator(
            mode="number+delta",
            value=round(data['performance_metrics']['spike'], 2),
            title={'text': titles[3], 'font': {'size': 12, 'color': secondary_color, 'family': 'Arial'}},
            delta={'reference': 8.0, 'relative': True, 'font': {'color': primary_color}},
            number={'font': {'size': 18, 'color': secondary_color}}
        ), row=1, col=4)
        
        # Row 2 - Engagement Metrics
        emotional_impact = (data['emotional_profile']['happiness'] + 
                          data['emotional_profile']['surprise'] + 
                          abs(data['emotional_profile']['sadness']) * 0.5) / 2
        
        fig.add_trace(go.Indicator(
            mode="gauge+number",
            value=round(emotional_impact, 1),
            title={'text': titles[4], 'font': {'size': 12, 'color': primary_color, 'family': 'Arial'}},
            gauge={
                'axis': {'range': [0, 50], 'tickcolor': primary_color},
                'bar': {'color': primary_color, 'thickness': 0.7},
                'bgcolor': "rgba(233, 30, 99, 0.1)",
                'steps': [
                    {'range': [0, 15], 'color': '#FCE4EC'},
                    {'range': [15, 30], 'color': '#F8BBD9'},
                    {'range': [30, 50], 'color': accent_color}
                ]
            },
            number={'font': {'size': 16, 'color': primary_color}, 'suffix': '%'}
        ), row=2, col=1)
        
        fig.add_trace(go.Indicator(
            mode="number+delta",
            value=round(data['performance_metrics']['dynamism_index'], 1),
            title={'text': titles[5], 'font': {'size': 12, 'color': secondary_color, 'family': 'Arial'}},
            delta={'reference': 15.0, 'relative': True},
            number={'font': {'size': 18, 'color': secondary_color}}
        ), row=2, col=2)
        
        fig.add_trace(go.Indicator(
            mode="number",
            value=round(data['performance_metrics']['decimal_star'], 1),
            title={'text': titles[6], 'font': {'size': 12, 'color': primary_color, 'family': 'Arial'}},
            number={'suffix': " ⭐", 'font': {'size': 18, 'color': primary_color}}
        ), row=2, col=3)
        
        fig.add_trace(go.Indicator(
            mode="number",
            value=int(data['duration']),
            title={'text': titles[7], 'font': {'size': 12, 'color': secondary_color, 'family': 'Arial'}},
            number={'suffix': " sec", 'font': {'size': 18, 'color': secondary_color}}
        ), row=2, col=4)
        
        # Row 3 - Context Metrics
        fig.add_trace(go.Indicator(
            mode="number",
            value=int(data['sample_info']['count_completes']),
            title={'text': titles[8], 'font': {'size': 12, 'color': primary_color, 'family': 'Arial'}},
            number={'font': {'size': 16, 'color': primary_color}}
        ), row=3, col=1)
        
        market_scores = {'US': 10, 'UK': 9, 'DE': 8, 'FR': 8, 'JP': 9, 'AU': 7, 'IN': 7, 'BR': 6, 'CA': 8}
        market_score = market_scores.get(data['market'], 5)
        
        fig.add_trace(go.Indicator(
            mode="number",
            value=market_score,
            title={'text': titles[9], 'font': {'size': 12, 'color': secondary_color, 'family': 'Arial'}},
            number={'suffix': "/10", 'font': {'size': 16, 'color': secondary_color}}
        ), row=3, col=2)
        
        premium_brands = ['Apple', 'Google', 'Microsoft', 'Amazon', 'Tesla']
        brand_score = 10 if data['brand'] in premium_brands else 7
        
        fig.add_trace(go.Indicator(
            mode="number",
            value=brand_score,
            title={'text': titles[10], 'font': {'size': 12, 'color': primary_color, 'family': 'Arial'}},
            number={'suffix': "/10", 'font': {'size': 16, 'color': primary_color}}
        ), row=3, col=3)
        
        tech_categories = ['Technology', 'E-commerce']
        category_score = 9 if data['category'] in tech_categories else 7
        
        fig.add_trace(go.Indicator(
            mode="number",
            value=category_score,
            title={'text': titles[11], 'font': {'size': 12, 'color': secondary_color, 'family': 'Arial'}},
            number={'suffix': "/10", 'font': {'size': 16, 'color': secondary_color}}
        ), row=3, col=4)
        
        fig.update_layout(
            height=900,
            showlegend=False,
            paper_bgcolor='white',
            plot_bgcolor='rgba(248, 187, 217, 0.05)',
            font={'color': '#333', 'family': 'Arial', 'size': 12},
            title={
                'text': f'🎯 Executive Performance Dashboard<br><span style="font-size:16px;color:#666">{data["title"]} - {data["brand"]}</span>',
                'x': 0.5,
                'font': {'size': 24, 'color': primary_color, 'family': 'Arial'},
                'y': 0.95
            },
            margin=dict(t=120, b=50, l=50, r=50)
        )
        
        return fig
    
    @staticmethod
    def create_emotional_radar_chart(data: Dict) -> go.Figure:
        """Create emotional profile radar chart"""
        emotions = list(data['emotional_profile'].keys())
        values = list(data['emotional_profile'].values())
        
        emotion_labels = {
            'contempt': '😤 Contempt', 'disgust': '🤢 Disgust', 'anger': '😡 Anger',
            'fear': '😨 Fear', 'sadness': '😢 Sadness', 'neutral': '😐 Neutral',
            'happiness': '😊 Happiness', 'surprise': '😲 Surprise'
        }
        
        labeled_emotions = [emotion_labels.get(emotion, emotion.title()) for emotion in emotions]
        
        fig = go.Figure()
        
        fig.add_trace(go.Scatterpolar(
            r=values + [values[0]],
            theta=labeled_emotions + [labeled_emotions[0]],
            fill='toself',
            name='Campaign Profile',
            line=dict(color='#E91E63', width=3),
            fillcolor='rgba(233, 30, 99, 0.3)',
            marker=dict(size=8, color='#C2185B', symbol='circle')
        ))
        
        benchmark_values = {'contempt': 3, 'disgust': 2, 'anger': 4, 'fear': 6, 'sadness': 8, 'neutral': 35, 'happiness': 28, 'surprise': 14}
        benchmark = [benchmark_values.get(emotion, 10) for emotion in emotions]
        
        fig.add_trace(go.Scatterpolar(
            r=benchmark + [benchmark[0]],
            theta=labeled_emotions + [labeled_emotions[0]],
            fill='toself',
            name='Industry Benchmark',
            line=dict(color='#666', width=2, dash='dot'),
            fillcolor='rgba(102, 102, 102, 0.2)',
            marker=dict(size=6, color='#444', symbol='diamond')
        ))
        
        fig.update_layout(
            polar=dict(
                radialaxis=dict(
                    visible=True,
                    range=[0, max(max(values), max(benchmark)) + 5],
                    tickfont=dict(size=11, color='#666', family='Arial'),
                    gridcolor='rgba(233, 30, 99, 0.2)'
                ),
                angularaxis=dict(
                    tickfont=dict(size=12, color='#333', family='Arial'),
                    gridcolor='rgba(233, 30, 99, 0.2)'
                ),
                bgcolor='rgba(248, 187, 217, 0.05)'
            ),
            showlegend=True,
            title={
                'text': '🎭 Emotional Response Analysis',
                'x': 0.5,
                'font': {'size': 18, 'color': '#E91E63', 'family': 'Arial'}
            },
            height=650,
            paper_bgcolor='white',
            margin=dict(t=100, b=80, l=50, r=50)
        )
        
        return fig
    
    @staticmethod
    def create_performance_comparison(data: Dict) -> go.Figure:
        """Create performance metrics comparison"""
        metrics_data = data['performance_metrics']
        
        categories = ['EIA Score', 'Fluency', 'Spike', 'Dynamism', 'Stars']
        values = [
            min(metrics_data['eia'], 10),
            metrics_data['final_fluency'] / 10,
            min(metrics_data['spike'], 10),
            metrics_data['total_dynamism'] / 10,
            metrics_data['decimal_star'] * 2
        ]
        
        fig = go.Figure()
        
        fig.add_trace(go.Bar(
            x=categories,
            y=values,
            marker=dict(
                color=['#E91E63', '#C2185B', '#AD1457', '#F48FB1', '#F8BBD9'],
                opacity=0.85,
                line=dict(color='#880E4F', width=1.5)
            ),
            text=[f'{v:.1f}' for v in values],
            textposition='auto',
            textfont=dict(color='white', size=12, family='Arial')
        ))
        
        fig.update_layout(
            title={
                'text': '📊 Performance Metrics Comparison',
                'x': 0.5,
                'font': {'size': 18, 'color': '#E91E63', 'family': 'Arial'}
            },
            xaxis_title='Metrics',
            yaxis_title='Normalized Score (0-10)',
            height=550,
            paper_bgcolor='white',
            plot_bgcolor='rgba(248, 187, 217, 0.05)',
            font=dict(family='Arial', size=11)
        )
        
        return fig
    
    @staticmethod
    def create_left_brain_chart(data: Dict) -> go.Figure:
        """Create left-brain (logical) metrics chart"""
        metrics_data = data['performance_metrics']
        
        categories = ['EIA', 'Fluency', 'Dynamism Index']
        values = [
            min(metrics_data['eia'], 10),
            metrics_data['final_fluency'] / 10,
            min(metrics_data['dynamism_index'] / 10, 10)
        ]
        
        fig = go.Figure()
        
        fig.add_trace(go.Bar(
            x=categories,
            y=values,
            marker=dict(
                color='#C2185B',
                opacity=0.85,
                line=dict(color='#880E4F', width=1.5)
            ),
            text=[f'{v:.1f}' for v in values],
            textposition='auto',
            textfont=dict(color='white', size=12, family='Arial')
        ))
        
        fig.update_layout(
            title={
                'text': '🧠 Left-Brain Metrics (Logical)',
                'x': 0.5,
                'font': {'size': 18, 'color': '#C2185B', 'family': 'Arial'}
            },
            xaxis_title='Logical Metrics',
            yaxis_title='Normalized Score (0-10)',
            height=450,
            paper_bgcolor='white',
            plot_bgcolor='rgba(248, 187, 217, 0.05)',
            font=dict(family='Arial', size=11)
        )
        
        return fig
    
    @staticmethod
    def create_right_brain_chart(data: Dict) -> go.Figure:
        """Create right-brain (emotional) metrics chart"""
        emotional_data = data['emotional_profile']
        
        categories = ['Happiness', 'Surprise', 'Sadness']
        values = [
            emotional_data['happiness'],
            emotional_data['surprise'],
            emotional_data['sadness']
        ]
        
        fig = go.Figure()
        
        fig.add_trace(go.Bar(
            x=categories,
            y=values,
            marker=dict(
                color='#E91E63',
                opacity=0.85,
                line=dict(color='#880E4F', width=1.5)
            ),
            text=[f'{v:.1f}%' for v in values],
            textposition='auto',
            textfont=dict(color='white', size=12, family='Arial')
        ))
        
        fig.update_layout(
            title={
                'text': '🎨 Right-Brain Metrics (Emotional)',
                'x': 0.5,
                'font': {'size': 18, 'color': '#E91E63', 'family': 'Arial'}
            },
            xaxis_title='Emotional Metrics',
            yaxis_title='Percentage (%)',
            height=450,
            paper_bgcolor='white',
            plot_bgcolor='rgba(248, 187, 217, 0.05)',
            font=dict(family='Arial', size=11)
        )
        
        return fig

class RecommendationsGenerator:
    """Generates optimization recommendations with detailed explanations"""
    
    def __init__(self):
        self.claude_client = get_claude_client()
        self.raw_response = None  # Store raw Claude response
    
    def generate_recommendations(self, data: Dict) -> List[Dict]:
        """Generate optimization recommendations with explanations"""
        if self.claude_client:
            try:
                prompt = self._create_optimization_prompt(data)
                message = self.claude_client.messages.create(
                    model="claude-3-5-sonnet-20241022",
                    max_tokens=1000,
                    temperature=0.7,
                    messages=[{"role": "user", "content": prompt}]
                )
                self.raw_response = message.content[0].text
                return self._parse_recommendations(self.raw_response, data)
            except Exception as e:
                print(f"Claude API error: {e}")
                self.raw_response = None
                return self._generate_fallback_recommendations(data)
        else:
            self.raw_response = None
            return self._generate_fallback_recommendations(data)
    
    def _create_optimization_prompt(self, data: Dict) -> str:
        """Create prompt for optimization recommendations"""
        return f"""Provide optimization recommendations for the following campaign data:
Metrics: {json.dumps(data['performance_metrics'])}
Emotional Profile: {json.dumps(data['emotional_profile'])}
Market: {data['market']}
Duration: {data['duration']} seconds
Category: {data['category']}
Brand: {data['brand']}
Provide a JSON response with:
{{
  "recommendations": [
    {{
      "priority": "High|Medium|Low",
      "action": "Specific actionable recommendation",
      "impact": "Expected performance lift",
      "explanation": "Detailed explanation of why this recommendation is made and how to implement it"
    }}
  ]
}}"""
    
    def _parse_recommendations(self, response_text: str, data: Dict) -> List[Dict]:
        """Parse Claude's recommendations response"""
        try:
            response = json.loads(response_text)
            recommendations = response.get('recommendations', [])
            # Ensure explanation field exists; if not, add a default one
            for rec in recommendations:
                if 'explanation' not in rec:
                    rec['explanation'] = f"This recommendation aims to improve engagement by focusing on key metrics like {data['performance_metrics']['eia']:.1f} EIA and emotional response such as {data['emotional_profile']['happiness']:.1f}% happiness."
            return recommendations
        except json.JSONDecodeError:
            return self._generate_fallback_recommendations(data)
    
    def _generate_fallback_recommendations(self, data: Dict) -> List[Dict]:
        """Generate detailed fallback recommendations based on campaign data"""
        metrics = data['performance_metrics']
        emotions = data['emotional_profile']
        market = data['market']
        duration = data['duration']
        category = data['category']
        brand = data['brand']
        
        # Identify dominant emotion
        dominant_emotion = max(emotions.items(), key=lambda x: x[1])[0]
        dominant_value = emotions[dominant_emotion]
        
        # Determine weak metrics
        weak_metrics = []
        if metrics['eia'] < 5:
            weak_metrics.append(f"EIA ({metrics['eia']:.1f})")
        if metrics['final_fluency'] < 50:
            weak_metrics.append(f"Fluency ({metrics['final_fluency']:.1f})")
        if metrics['spike'] < 5:
            weak_metrics.append(f"Spike ({metrics['spike']:.1f})")
        
        # Market-specific adjustment
        market_strength = {'US': 10, 'UK': 9, 'DE': 8, 'FR': 8, 'JP': 9, 'AU': 7, 'IN': 7, 'BR': 6, 'CA': 8}.get(market, 5)
        
        recommendations = [
            {
                'priority': 'High',
                'action': f"Amplify {dominant_emotion} to drive engagement",
                'impact': '10-15% engagement lift',
                'explanation': f"The campaign's {dominant_emotion} score of {dominant_value:.1f}% is strong in the {market} market (market strength: {market_strength}/10). Emphasize {dominant_emotion} by featuring more scenes that showcase the {category} product's emotional benefits, such as customer testimonials or relatable scenarios for {brand}."
            },
            {
                'priority': 'Medium',
                'action': 'Adjust duration for optimal impact',
                'impact': '8-12% performance increase',
                'explanation': f"The current duration of {duration} seconds may {'be too long' if duration > 30 else 'limit impact'}. For {category} campaigns in {market}, {'shorten to 20-30 seconds to maintain attention' if duration > 30 else 'extend to 30-40 seconds to build narrative'}. Edit the ad to focus on key {brand} messages while maintaining emotional peaks."
            },
            {
                'priority': 'Medium',
                'action': f"Improve {' and '.join(weak_metrics) if weak_metrics else 'key performance metrics'}",
                'impact': '5-10% metric improvement',
                'explanation': f"{'Weak metrics include ' + ', '.join(weak_metrics) + '.' if weak_metrics else 'Metrics like EIA and Fluency can improve.'} For {brand}'s {category} campaign, enhance clarity by simplifying messaging and increasing visual cues. In {market}, consider A/B testing with varied calls-to-action to boost {'EIA and Spike' if weak_metrics else 'overall performance'}."
            },
            {
                'priority': 'Low',
                'action': 'Integrate market-specific cultural elements',
                'impact': '3-7% brand resonance increase',
                'explanation': f"To align with {market}'s cultural context, incorporate local references or trends relevant to {category}. For {brand}, this could mean using regional influencers or tailoring visuals to resonate with {market} audiences, enhancing emotional connection beyond the current {dominant_emotion} focus."
            }
        ]
        
        return recommendations

# Global variables
current_df = None
recommendations_generator = RecommendationsGenerator()
current_recommendations = []

def load_data(file):
    """Load and process uploaded data file"""
    global current_df
    
    if file is None:
        return "Please upload a CSV or Excel file.", gr.update(choices=[], value=None), "No data loaded"
    
    try:
        # Handle different types of file input
        file_path = None
        is_csv = False
        
        if isinstance(file, bytes):
            # Handle raw bytes
            with tempfile.NamedTemporaryFile(delete=False, suffix='.csv') as tmp:
                tmp.write(file)
                file_path = tmp.name
                is_csv = True  # Assume CSV for raw bytes
        elif hasattr(file, 'name'):
            # Handle file object with name attribute
            file_path = file.name
            is_csv = file_path.endswith('.csv')
        elif isinstance(file, gr.FileData):
            # Handle Gradio FileData object
            file_path = file.path
            is_csv = file_path.endswith('.csv')
        else:
            raise ValueError("Unsupported file input type")
        
        current_df = DataProcessor.load_and_clean_data(file_path)
        
        # Clean up temporary file if created
        if isinstance(file, bytes):
            os.unlink(file_path)
        
        if current_df is None:
            return "Error: Could not load data. Please check file format and required columns (CreativeID, Title, ReportingBrand).", gr.update(choices=[], value=None), "Load failed"
        
        creative_ids = DataProcessor.get_creative_ids_list(current_df)
        
        return (
            f"✅ Successfully loaded {len(current_df)} campaigns from {len(creative_ids)} Creative IDs",
            gr.update(choices=creative_ids, value=creative_ids[0] if creative_ids else None),
            f"Data loaded: {len(current_df)} rows, {len(current_df.columns)} columns"
        )
        
    except Exception as e:
        return f"Error loading file: {str(e)}", gr.update(choices=[], value=None), "Load error"

def analyze_campaign(creative_id):
    """Analyze selected campaign and generate charts, recommendations, and insights"""
    global current_df, recommendations_generator, current_recommendations
    
    if current_df is None:
        return "Please load data first.", None, None, None, None, None, False, "<p>No recommendations available.</p>"
    
    if not creative_id:
        return "Please select a Creative ID.", None, None, None, None, None, False, "<p>No recommendations available.</p>"
    
    try:
        campaign_data = DataProcessor.get_campaign_data(current_df, creative_id)
        if campaign_data is None:
            return "Campaign not found.", None, None, None, None, None, False, "<p>No recommendations available.</p>"
        
        dashboard_chart = ChartGenerator.create_executive_dashboard(campaign_data)
        emotional_chart = ChartGenerator.create_emotional_radar_chart(campaign_data)
        performance_chart = ChartGenerator.create_performance_comparison(campaign_data)
        left_brain_chart = ChartGenerator.create_left_brain_chart(campaign_data)
        right_brain_chart = ChartGenerator.create_right_brain_chart(campaign_data)
        current_recommendations = recommendations_generator.generate_recommendations(campaign_data)
        recommendations_html = show_recommendations(campaign_data['brand'], campaign_data['performance_metrics']['decimal_star'])
        
        return (
            f"✅ Analysis complete for {campaign_data['title']} ({creative_id})",
            dashboard_chart,
            emotional_chart,
            performance_chart,
            left_brain_chart,
            right_brain_chart,
            True,  # Enable recommendation buttons
            recommendations_html
        )
    except Exception as e:
        return f"Error analyzing campaign: {str(e)}", None, None, None, None, None, False, "<p>Error generating recommendations.</p>"

def show_recommendations(brand: str, decimal_star: float) -> str:
    """Generate HTML for recommendations and AI insights"""
    global current_recommendations
    
    if not current_recommendations:
        return "<p>No recommendations available.</p>"
    
    # Ensure we have exactly 4 recommendations to match the image layout
    insights = []
    for i in range(4):
        if i < len(current_recommendations):
            rec = current_recommendations[i]
            priority_star = "★" * (3 if rec['priority'] == "High" else 2 if rec['priority'] == "Medium" else 1)
            insights.append({
                'priority': priority_star,
                'title': rec['action'].upper(),
                'description': rec['explanation']
            })
        else:
            # Fallback if fewer than 4 recommendations
            insights.append({
                'priority': "★",
                'title': "CONTINUE MONITORING PERFORMANCE",
                'description': "Keep tracking key metrics to identify further optimization opportunities."
            })
    
    html = """
    <div style='padding: 20px; background: linear-gradient(45deg, #E91E63, #F8BBD9); border-radius: 15px;'>
        <h2 style='color: white; font-family: Arial; text-align: center; margin-bottom: 20px; font-size: 24px; text-transform: uppercase;'>Recommendations</h2>
        <p style='color: white; font-family: Arial; text-align: center; margin-bottom: 30px; font-size: 16px;'>
            At {decimal_star:.1f} Stars, this campaign demonstrates modest long-term growth potential for {brand} but consider the following optimizations for now and in the future.
        </p>
        <div style='display: grid; grid-template-columns: repeat(2, 1fr); gap: 20px;'>
    """.format(decimal_star=decimal_star, brand=brand)
    
    for idx, insight in enumerate(insights):
        html += f"""
        <div class='insight-card'>
            <div style='display: flex; align-items: center; margin-bottom: 10px;'>
                <span style='color: #C2185B; font-size: 20px; margin-right: 10px;'>{insight['priority']}</span>
                <h3 style='color: #333; font-family: Arial; font-size: 16px; font-weight: bold; margin: 0; text-transform: uppercase;'>{insight['title']}</h3>
            </div>
            <p style='color: #333; font-family: Arial; font-size: 14px; line-height: 1.5;'>{insight['description']}</p>
        </div>
        """
    
    html += """
        </div>
        <p style='color: white; font-family: Arial; text-align: right; font-size: 12px; margin-top: 20px;'>© System1 Group PLC</p>
    </div>
    """
    return html

def create_sample_data():
    """Create sample dataset for demonstration"""
    global current_df
    
    sample_data = {
        'CreativeID': ['CR35383', 'CR82756', 'CR10436', 'CR90002', 'CR43967'],
        'Title': [
            'Microsoft - Premium Smartphones Campaign',
            'Google - Classic Mobile Plans Campaign', 
            'Amazon - New Mobile Plans Campaign',
            'Microsoft - Premium Soft Drinks Campaign',
            'Starbucks - New Soft Drinks Campaign'
        ],
        'ReportingBrand': ['Microsoft', 'Google', 'Amazon', 'Microsoft', 'Starbucks'],
        'ReportingAdvertiser': ['Elite Advertising', 'Dynamic Advertising Group', 'Elite Advertising', 'Platinum Marketing', 'Elite Advertising'],
        'ReportingSubCategory': ['Technology', 'Telecommunications', 'Telecommunications', 'Beverages', 'Beverages'],
        'Market': ['US', 'FR', 'JP', 'US', 'IN'],
        'Duration': [48, 20, 23, 72, 27],
        'EIA': [6.6, 3.7, 7.8, 3.0, 2.9],
        'DecimalStar': [3.4, 4.5, 2.8, 1.5, 3.0],
        'FinalFluency': [18.7, 0.9, 18.2, 0.2, 96.1],
        'FastFluency': [73.8, 96.5, 91.9, 97.0, 98.1],
        'Spike': [7.33, 3.5, 1.84, 8.05, 0.15],
        'OEI': [53.1, 34.2, 37.3, 8.3, 52.0],
        'TotalDynamism': [73.5, 38.6, 17.6, 11.1, 58.8],
        'DynamismIndex': [66.7, 56.1, 107.2, 136.3, 126.3],
        'Happiness (%)': [29.5, 17.0, 25.6, 31.0, 29.4],
        'Surprise (%)': [12.1, 7.9, 9.8, 6.0, 7.9],
        'Neutral (%)': [24.8, 49.9, 21.6, 38.6, 34.6],
        'Sadness (%)': [10.7, 0.2, 7.3, 6.8, 0.5],
        'Fear (%)': [3.7, 1.8, 4.5, 3.3, 3.6],
        'Anger (%)': [14.7, 12.3, 12.6, 3.4, 6.4],
        'Contempt (%)': [3.6, 6.9, 10.8, 8.0, 13.1],
        'Disgust (%)': [0.9, 4.0, 7.8, 2.9, 4.5],
        'CountCompletes': [196, 331, 183, 168, 459],
        'SampleDescription': ['Sample of 196 respondents from UK', 'Sample of 331 respondents from UK', 'Sample of 183 respondents from CN', 'Sample of 168 respondents from DE', 'Sample of 459 respondents from AU'],
        'SampleDetails': ['Mixed respondents, ages 27-42', 'Mixed respondents, ages 20-36', 'Mixed respondents, ages 26-37', 'Male respondents, ages 20-38', 'Mixed respondents, ages 27-34'],
        'DisplayType': ['Video', 'Banner', 'Interactive', 'Image', 'Image'],
        'SurveyType': ['Product Testing', 'Ad Effectiveness', 'Consumer Feedback', 'Product Testing', 'Brand Awareness'],
        'StimuliType': ['Radio Spot', 'Radio Spot', 'Commercial', 'Billboard', 'Billboard'],
        'AdvertMedium': ['Online', 'Online', 'Online', 'Television', 'Television'],
        'AdvertFormat': ['Social media post', 'Banner ad', 'Banner ad', 'Banner ad', 'Pre-roll video']
    }
    
    current_df = pd.DataFrame(sample_data)
    current_df = DataProcessor.calculate_creativity_scores(current_df)
    creative_ids = DataProcessor.get_creative_ids_list(current_df)
    
    return (
        f"✅ Sample data loaded: {len(current_df)} campaigns",
        gr.update(choices=creative_ids, value=creative_ids[0] if creative_ids else None),
        f"Sample data ready for analysis"
    )

# Custom CSS for enhanced UI
custom_css = """
.gradio-container {
    background: linear-gradient(135deg, #fce4ec 0%, #ffffff 100%);
    font-family: 'Arial', sans-serif;
}

.gr-button {
    background: linear-gradient(45deg, #C2185B, #E91E63) !important;
    color: white !important;
    border: none !important;
    border-radius: 10px !important;
    font-weight: bold !important;
    transition: all 0.3s ease !important;
    padding: 12px 24px !important;
}

.gr-button:hover {
    background: linear-gradient(45deg, #AD1457, #C2185B) !important;
    transform: translateY(-2px) !important;
    box-shadow: 0 6px 20px rgba(194, 24, 91, 0.3) !important;
}

.gr-form {
    background: white !important;
    border-radius: 15px !important;
    box-shadow: 0 8px 25px rgba(194, 24, 91, 0.1) !important;
    border: 2px solid #F8BBD9 !important;
    padding: 20px !important;
}

.gr-panel {
    background: white !important;
    border-radius: 12px !important;
    border: 1px solid #F8BBD9 !important;
    padding: 15px !important;
}

h1 {
    color: white !important;
    font-weight: bold !important;
    text-shadow: 1px 1px 2px rgba(0,0,0,0.2) !important;
}

h2 {
    color: #C2185B !important;
    font-weight: bold !important;
    margin-bottom: 15px !important;
}

.gr-dropdown {
    border: 2px solid #F8BBD9 !important;
    border-radius: 10px !important;
    padding: 8px !important;
}

.gr-file {
    border: 2px dashed #E91E63 !important;
    border-radius: 12px !important;
    background: #FCE4EC !important;
    padding: 15px !important;
}

.gr-textbox {
    border: 2px solid #F8BBD9 !important;
    border-radius: 10px !important;
    padding: 10px !important;
    background: #FCE4EC !important;
}

.gr-html {
    border-radius: 12px !important;
    border: 1px solid #F8BBD9 !important;
    padding: 20px !important;
    background: white !important;
}

.insight-card {
    background: white !important;
    border-radius: 10px !important;
    padding: 15px !important;
    box-shadow: 0 4px 15px rgba(0, 0, 0, 0.1) !important;
}
"""

def build_interface():
    with gr.Blocks(css=custom_css, title="🎯 Advanced Ad Campaign Analyzer") as demo:
        gr.HTML("""
        <div style="text-align: center; padding: 30px; background: linear-gradient(45deg, #C2185B, #E91E63); color: white; border-radius: 20px; margin-bottom: 25px;">
            <h1 style="margin: 0; font-size: 2.8em;"> Express Guidance Analyser </h1>
            <p style="margin: 15px 0 0 0; font-size: 1.3em; color: rgba(255,255,255,0.9);">
                AI-Powered Advertising Intelligence Platform
            </p>
        </div>
        """)
        
        with gr.Row():
            with gr.Column(scale=1):
                gr.HTML("<h2>📊 Data Management</h2>")
                
                file_input = gr.File(
                    label="Upload Campaign Data (CSV/Excel)",
                    file_types=[".csv", ".xlsx", ".xls"],
                    type="binary"
                )
                load_button = gr.Button("🚀 Load Data", variant="primary")
                sample_button = gr.Button("📋 Use Sample Data", variant="primary")
                
                load_status = gr.Textbox(label="Load Status", interactive=False)
                
                creative_dropdown = gr.Dropdown(
                    label="Select Creative ID for Analysis",
                    choices=[],
                    interactive=True
                )
                
                analyze_button = gr.Button("🔍 Analyze Campaign", variant="primary")
            
            with gr.Column(scale=3):
                gr.HTML("<h2>📈 Analysis Results</h2>")
                
                analysis_status = gr.Textbox(label="Analysis Status", interactive=False)
                
                with gr.Tabs():
                    with gr.TabItem("📊 Executive Dashboard"):
                        dashboard_plot = gr.Plot(label="Performance Dashboard")
                    
                    with gr.TabItem("🎭 Emotional Analysis"):
                        emotional_plot = gr.Plot(label="Emotional Profile")
                    
                    with gr.TabItem("📈 Performance Metrics"):
                        performance_plot = gr.Plot(label="Performance Comparison")
                    
                    with gr.TabItem("🧠 Left-Brain Metrics"):
                        left_brain_plot = gr.Plot(label="Logical Metrics")
                    
                    with gr.TabItem("🎨 Right-Brain Metrics"):
                        right_brain_plot = gr.Plot(label="Emotional Metrics")
                
                gr.HTML("<h2>🚀 AI Insights & Recommendations</h2>")
                insights_output = gr.HTML(label="AI Insights & Recommendations")
        
        load_button.click(
            fn=load_data,
            inputs=file_input,
            outputs=[load_status, creative_dropdown, analysis_status]
        )
        
        sample_button.click(
            fn=create_sample_data,
            inputs=None,
            outputs=[load_status, creative_dropdown, analysis_status]
        )
        
        analyze_button.click(
            fn=analyze_campaign,
            inputs=creative_dropdown,
            outputs=[
                analysis_status,
                dashboard_plot,
                emotional_plot,
                performance_plot,
                left_brain_plot,
                right_brain_plot,
                gr.State(),  # State to enable/disable buttons
                insights_output
            ]
        )
        
        return demo

if __name__ == "__main__":
    demo = build_interface()
    demo.launch()