# report_generator.py

from typing import Dict, List
import numpy as np
from collections import defaultdict
from datetime import datetime

class ReportGenerator:
    """
    Generates textual and HTML reports from evaluation results
    """

    def generate_batch_report(self, results: List[Dict]) -> str:
        """
        Generates a comprehensive text report for a batch evaluation.
        
        Args:
            results: A list of evaluation result dictionaries.
            
        Returns:
            A formatted string containing the batch evaluation report.
        """
        if not results:
            return "No evaluation results to report."

        num_evals = len(results)
        agent_names = list(set(r['agent_name'] for r in results))
        num_agents = len(agent_names)

        # Aggregate scores
        overall_scores = [r['scores']['overall_score'] for r in results]
        metric_scores = defaultdict(list)
        agent_overall_scores = defaultdict(list)

        for res in results:
            agent_overall_scores[res['agent_name']].append(res['scores']['overall_score'])
            for metric, score in res['scores'].items():
                if metric not in ['eval_id', 'timestamp', 'task_type']:
                    metric_scores[metric].append(score)

        # Calculate agent averages
        agent_avg_scores = {agent: np.mean(scores) for agent, scores in agent_overall_scores.items()}
        top_agent = max(agent_avg_scores, key=agent_avg_scores.get)
        bottom_agent = min(agent_avg_scores, key=agent_avg_scores.get)

        # Build report string
        report = []
        report.append("="*50)
        report.append("    AetherScore - Batch Evaluation Report")
        report.append("="*50)
        report.append(f"Report generated on: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n")
        
        report.append("--- Summary ---")
        report.append(f"Total Evaluations: {num_evals}")
        report.append(f"Number of Agents: {num_agents}")
        report.append(f"Overall Average Score: {np.mean(overall_scores):.3f}\n")

        report.append("--- Agent Performance ---")
        report.append(f"Top Performing Agent: {top_agent} (Avg Score: {agent_avg_scores[top_agent]:.3f})")
        report.append(f"Agent with most room for improvement: {bottom_agent} (Avg Score: {agent_avg_scores[bottom_agent]:.3f})\n")

        report.append("--- Metric Breakdown (Average Scores) ---")
        for metric, scores in metric_scores.items():
            metric_name = metric.replace('_', ' ').title()
            report.append(f"- {metric_name:<25}: {np.mean(scores):.3f}")
        
        report.append("\n" + "="*50)
        
        return "\n".join(report)

    def generate_comparison_report(
        self,
        agent1_results: List[Dict],
        agent2_results: List[Dict]
    ) -> str:
        """
        Generates a text report comparing two agents.
        
        Args:
            agent1_results: Evaluation results for the first agent.
            agent2_results: Evaluation results for the second agent.
            
        Returns:
            A formatted string comparing the two agents.
        """
        if not agent1_results or not agent2_results:
            return "Insufficient data for comparison. Please provide results for both agents."

        agent1_name = agent1_results[0].get('agent_name', 'Agent 1')
        agent2_name = agent2_results[0].get('agent_name', 'Agent 2')

        # Calculate average scores for each agent
        metrics = ['overall_score', 'instruction_following', 'hallucination_score', 
                   'assumption_control', 'coherence', 'accuracy']
        
        avg_scores1 = {m: np.mean([r['scores'].get(m, 0) for r in agent1_results]) for m in metrics}
        avg_scores2 = {m: np.mean([r['scores'].get(m, 0) for r in agent2_results]) for m in metrics}
        
        # Build report string
        report = []
        report.append("="*60)
        report.append(f"    Agent Comparison Report: {agent1_name} vs. {agent2_name}")
        report.append("="*60)
        report.append(f"Report generated on: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}\n")
        
        # Overall Winner
        winner = agent1_name if avg_scores1['overall_score'] > avg_scores2['overall_score'] else agent2_name
        report.append("--- Overall Performance ---")
        report.append(f"🏆 Winner: {winner}")
        report.append(f"{agent1_name} Avg Overall Score: {avg_scores1['overall_score']:.3f}")
        report.append(f"{agent2_name} Avg Overall Score: {avg_scores2['overall_score']:.3f}\n")

        report.append("--- Detailed Metric Comparison ---")
        header = f"{'Metric':<25} | {agent1_name:<10} | {agent2_name:<10} | {'Delta':<8} | {'Winner'}"
        report.append(header)
        report.append("-"*len(header))

        for metric in metrics:
            s1 = avg_scores1[metric]
            s2 = avg_scores2[metric]
            delta = s2 - s1
            metric_winner = agent1_name if s1 > s2 else agent2_name if s2 > s1 else "Tie"
            metric_name = metric.replace('_', ' ').title()
            
            report.append(f"{metric_name:<25} | {s1:<10.3f} | {s2:<10.3f} | {delta:<+8.3f} | {metric_winner}")

        report.append("\n" + "="*60)
        
        return "\n".join(report)

    def generate_html_report(self, results_data: List[Dict]) -> str:
        """
        Generates a basic HTML report from evaluation results.
        
        Args:
            results_data: A list of evaluation result dictionaries.
            
        Returns:
            A string containing a full HTML report.
        """
        report_str = self.generate_batch_report(results_data)
        
        # Basic HTML template
        html_template = f"""
        <!DOCTYPE html>
        <html lang="en">
        <head>
            <meta charset="UTF-8">
            <meta name="viewport" content="width=device-width, initial-scale=1.0">
            <title>AetherScore Evaluation Report</title>
            <style>
                body {{ font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif; line-height: 1.6; color: #333; }}
                .container {{ max-width: 800px; margin: auto; padding: 20px; }}
                h1 {{ color: #4a4a4a; }}
                pre {{ background: #f4f4f4; padding: 15px; border-radius: 5px; white-space: pre-wrap; }}
            </style>
        </head>
        <body>
            <div class="container">
                <h1>AetherScore Evaluation Report</h1>
                <p>This report contains a summary of the batch evaluation results.</p>
                <pre>{report_str}</pre>
                <p><em>Note: This is a text-based summary. For interactive visualizations, please use the AetherScore dashboard.</em></p>
            </div>
        </body>
        </html>
        """
        return html_template