modius_performance/visualization/roi_charts.py

"""
ROI chart implementations.
"""
import plotly.graph_objects as go
import pandas as pd
import logging
from datetime import datetime
from typing import Dict, Any, Optional, Tuple

from ..config.constants import DATE_RANGES, Y_AXIS_RANGES, FILE_PATHS
from .base_chart import BaseChart

logger = logging.getLogger(__name__)


class ROIChart(BaseChart):
    """Chart for ROI visualizations."""
    
    def create_chart(self, df: pd.DataFrame, **kwargs) -> go.Figure:
        """Create ROI time series chart."""
        if df.empty:
            return self._create_empty_chart("No ROI data available")
        
        # Filter for ROI data only
        roi_data = df[df['metric_type'] == 'ROI'].copy()
        if roi_data.empty:
            return self._create_empty_chart("No ROI data available")
        
        # ROI data is already in percentage format from corrected CSV
        # No conversion needed
        
        # Apply daily median aggregation to reduce outliers
        roi_data = self.data_processor.aggregate_daily_medians(roi_data, ['roi'])
        if roi_data.empty:
            return self._create_empty_chart("No ROI data available after aggregation")
        
        # Apply high APR filtering to ROI data (need to get APR data first to determine what to filter)
        # Fetch APR data to determine which agent-timestamp combinations to filter
        apr_df, _ = self.data_processor.fetch_apr_data_from_db()
        if not apr_df.empty:
            # Apply same processing to APR data to get the filtering reference
            apr_processed = apr_df[apr_df['metric_type'] == 'APR'].copy()
            apr_processed = self.data_processor.aggregate_daily_medians(apr_processed, ['apr', 'adjusted_apr'])
            
            # Apply high APR filtering to both datasets
            _, roi_data = self.data_processor.filter_high_apr_values(apr_processed, roi_data)
        
        if roi_data.empty:
            return self._create_empty_chart("No ROI data available after high APR filtering")
        
        # Save processed ROI data for verification (after all processing steps)
        processed_csv_path = self.data_processor.save_to_csv(roi_data, FILE_PATHS['roi_processed_csv'])
        if processed_csv_path:
            logger.info(f"Saved processed ROI data to {processed_csv_path} for verification")
            logger.info(f"Processed ROI data contains {len(roi_data)} rows after agent exclusion, daily median aggregation, and high APR filtering")
        
        # Filter outliers (disabled but keeping for compatibility)
        roi_data = self._filter_outliers(roi_data, 'roi')
        
        # Get time range
        min_time = roi_data['timestamp'].min()
        max_time = roi_data['timestamp'].max()
        x_start_date = min_time  # Use actual start date for ROI
        
        # Calculate average runtime for title
        fixed_start_date = DATE_RANGES['feb_start']
        agent_runtimes = {}
        for agent_id in roi_data['agent_id'].unique():
            agent_data = roi_data[roi_data['agent_id'] == agent_id]
            agent_name = agent_data['agent_name'].iloc[0]
            last_report = agent_data['timestamp'].max()
            runtime_days = (last_report - fixed_start_date).total_seconds() / (24 * 3600)
            agent_runtimes[agent_id] = {
                'agent_name': agent_name,
                'last_report': last_report,
                'runtime_days': runtime_days
            }
        
        avg_runtime = sum(data['runtime_days'] for data in agent_runtimes.values()) / len(agent_runtimes) if agent_runtimes else 0
        
        # Create figure
        fig = self._create_base_figure()
        
        # Add background shapes
        y_range = Y_AXIS_RANGES['roi']
        self._add_background_shapes(fig, min_time, max_time, y_range['min'], y_range['max'])
        self._add_zero_line(fig, min_time, max_time)
        
        # Calculate moving averages
        avg_roi_data = self._calculate_moving_average(roi_data, 'roi')
        
        # Add individual agent data points
        unique_agents = roi_data['agent_name'].unique()
        color_map = self._get_color_map(unique_agents)
        self._add_agent_data_points(fig, roi_data, 'roi', color_map)
        
        # Add ROI moving average line
        self._add_moving_average_line(
            fig, avg_roi_data, 'roi', 
            'Average ROI (3d window)', 
            self.colors['roi'], 
            width=2
        )
        
        # Update layout and axes
        title = f"Modius Agents ROI (over avg. {avg_runtime:.1f} days runtime)"
        self._update_layout(
            fig, 
            title=title,
            y_axis_title="ROI [%]",
            y_range=[y_range['min'], y_range['max']]
        )
        
        # Use auto-range for both x-axis and y-axis to show all available data
        self._update_axes(
            fig,
            x_range=None,  # Let plotly auto-determine the best x-axis range
            y_auto=True    # Let plotly auto-determine the best y-axis range
        )
        
        # Save chart
        self._save_chart(
            fig, 
            FILE_PATHS['roi_graph_html'], 
            FILE_PATHS['roi_graph_png']
        )
        
        return fig


def generate_roi_visualizations(data_processor=None) -> Tuple[go.Figure, Optional[str]]:
    """Generate ROI visualizations."""
    from ..data.data_processor import DataProcessor
    
    if data_processor is None:
        data_processor = DataProcessor()
    
    # Fetch data
    _, roi_df = data_processor.fetch_apr_data_from_db()
    
    # Create chart
    roi_chart = ROIChart(data_processor)
    fig, csv_path = roi_chart.generate_visualization(
        roi_df, 
        csv_filename=FILE_PATHS['roi_csv']
    )
    
    return fig, csv_path