Spaces:
Sleeping
Sleeping
File size: 14,745 Bytes
2f35ee2 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 |
#!/usr/bin/env python3
"""
OnCall.ai System - RAG vs Direct Latency Comparison Chart Generator
==================================================================
Compares RAG and Direct LLM system latency performance.
Reads statistics from latency_statistics_*.json and direct_llm_statistics_*.json
No LLM calls - pure data visualization.
Author: YanBo Chen
Date: 2025-08-05
"""
import json
import os
import sys
from typing import Dict, List, Any, Tuple
from datetime import datetime
from pathlib import Path
import glob
# Visualization imports
import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd
import numpy as np
class RAGvsDirectLatencyChartGenerator:
"""Generate RAG vs Direct latency comparison charts"""
def __init__(self):
"""Initialize chart generator"""
print("π Initializing RAG vs Direct Latency Chart Generator...")
# Set up professional chart style
plt.style.use('default')
sns.set_palette("husl")
# Define system colors
self.system_colors = {
'rag': '#1f77b4', # Blue
'direct': '#ff7f0e' # Orange
}
print("β
Chart Generator ready with professional medical styling")
def find_latest_statistics_files(self) -> Tuple[str, str]:
"""
Find the most recent RAG and Direct statistics files
Returns:
Tuple of (rag_file_path, direct_file_path)
"""
results_dir = Path(__file__).parent / "results"
# Find RAG statistics file
rag_pattern = str(results_dir / "latency_statistics_*.json")
rag_files = glob.glob(rag_pattern)
if not rag_files:
raise FileNotFoundError(f"No RAG latency statistics files found with pattern: {rag_pattern}")
latest_rag_file = max(rag_files, key=os.path.getmtime)
# Find Direct statistics file
direct_pattern = str(results_dir / "direct_llm_statistics_*.json")
direct_files = glob.glob(direct_pattern)
if not direct_files:
raise FileNotFoundError(f"No Direct LLM statistics files found with pattern: {direct_pattern}")
latest_direct_file = max(direct_files, key=os.path.getmtime)
print(f"π Found RAG statistics: {latest_rag_file}")
print(f"π Found Direct statistics: {latest_direct_file}")
return latest_rag_file, latest_direct_file
def load_statistics(self, rag_file: str, direct_file: str) -> Tuple[Dict, Dict]:
"""
Load statistics from both files
Args:
rag_file: Path to RAG statistics file
direct_file: Path to Direct statistics file
Returns:
Tuple of (rag_stats, direct_stats)
"""
print(f"π Loading RAG statistics from: {rag_file}")
with open(rag_file, 'r', encoding='utf-8') as f:
rag_stats = json.load(f)
print(f"π Loading Direct statistics from: {direct_file}")
with open(direct_file, 'r', encoding='utf-8') as f:
direct_stats = json.load(f)
return rag_stats, direct_stats
def generate_comparison_charts(self, rag_stats: Dict, direct_stats: Dict) -> str:
"""
Generate comprehensive RAG vs Direct latency comparison charts
Creates 4-panel comparison:
1. Category-wise latency comparison
2. Overall performance comparison
3. Target compliance comparison
4. Success rate comparison
"""
try:
# Create figure with subplots
fig, axes = plt.subplots(2, 2, figsize=(16, 12))
fig.suptitle('RAG vs Direct LLM - Latency Performance Comparison',
fontsize=16, fontweight='bold')
# Chart 1: Category-wise Latency Comparison
ax1 = axes[0, 0]
categories = ['diagnosis', 'treatment', 'mixed']
rag_latencies = []
direct_latencies = []
for category in categories:
rag_cat = rag_stats['category_results'].get(category, {})
direct_cat = direct_stats['category_results'].get(category, {})
rag_latencies.append(rag_cat.get('average_latency', 0))
direct_latencies.append(direct_cat.get('average_latency', 0))
x = np.arange(len(categories))
width = 0.35
bars1 = ax1.bar(x - width/2, rag_latencies, width, label='RAG',
color=self.system_colors['rag'], alpha=0.8)
bars2 = ax1.bar(x + width/2, direct_latencies, width, label='Direct LLM',
color=self.system_colors['direct'], alpha=0.8)
ax1.set_title('Latency by Category', fontweight='bold')
ax1.set_ylabel('Average Latency (seconds)')
ax1.set_xlabel('Query Category')
ax1.set_xticks(x)
ax1.set_xticklabels([cat.capitalize() for cat in categories])
ax1.legend()
ax1.grid(True, alpha=0.3)
# Add target line
ax1.axhline(y=60.0, color='red', linestyle='--', alpha=0.7, label='60s Target')
ax1.legend()
# Add value labels on bars
for bars in [bars1, bars2]:
for bar in bars:
height = bar.get_height()
if height > 0:
ax1.text(bar.get_x() + bar.get_width()/2., height + 1,
f'{height:.1f}s', ha='center', va='bottom', fontsize=9)
# Chart 2: Overall Performance Comparison
ax2 = axes[0, 1]
systems = ['RAG', 'Direct LLM']
overall_latencies = [
rag_stats['overall_results']['average_latency'],
direct_stats['overall_results']['average_latency']
]
bars = ax2.bar(systems, overall_latencies,
color=[self.system_colors['rag'], self.system_colors['direct']],
alpha=0.8)
ax2.set_title('Overall Average Latency', fontweight='bold')
ax2.set_ylabel('Average Latency (seconds)')
ax2.grid(True, alpha=0.3)
# Add target line
ax2.axhline(y=60.0, color='red', linestyle='--', alpha=0.7, label='60s Target')
ax2.legend()
# Add value labels
for bar, value in zip(bars, overall_latencies):
height = bar.get_height()
ax2.text(bar.get_x() + bar.get_width()/2., height + 1,
f'{value:.1f}s', ha='center', va='bottom', fontweight='bold')
# Chart 3: Target Compliance Comparison
ax3 = axes[1, 0]
rag_compliance = rag_stats['overall_results']['target_compliance'] * 100
direct_compliance = direct_stats['overall_results']['target_compliance'] * 100
compliance_data = [rag_compliance, direct_compliance]
bars = ax3.bar(systems, compliance_data,
color=[self.system_colors['rag'], self.system_colors['direct']],
alpha=0.8)
ax3.set_title('60s Target Compliance Rate', fontweight='bold')
ax3.set_ylabel('Compliance Rate (%)')
ax3.set_ylim(0, 105)
ax3.grid(True, alpha=0.3)
# Add target line at 100%
ax3.axhline(y=100.0, color='green', linestyle='--', alpha=0.7, label='100% Target')
ax3.legend()
# Add percentage labels
for bar, value in zip(bars, compliance_data):
height = bar.get_height()
ax3.text(bar.get_x() + bar.get_width()/2., height + 1,
f'{value:.1f}%', ha='center', va='bottom', fontweight='bold')
# Chart 4: Success Rate Comparison
ax4 = axes[1, 1]
rag_success_rate = rag_stats['overall_results']['successful_queries'] / rag_stats['overall_results']['total_queries'] * 100
direct_success_rate = direct_stats['overall_results']['successful_queries'] / direct_stats['overall_results']['total_queries'] * 100
success_data = [rag_success_rate, direct_success_rate]
bars = ax4.bar(systems, success_data,
color=[self.system_colors['rag'], self.system_colors['direct']],
alpha=0.8)
ax4.set_title('Query Success Rate', fontweight='bold')
ax4.set_ylabel('Success Rate (%)')
ax4.set_ylim(0, 105)
ax4.grid(True, alpha=0.3)
# Add target line at 100%
ax4.axhline(y=100.0, color='green', linestyle='--', alpha=0.7, label='100% Target')
ax4.legend()
# Add percentage labels
for bar, value in zip(bars, success_data):
height = bar.get_height()
ax4.text(bar.get_x() + bar.get_width()/2., height + 1,
f'{value:.1f}%', ha='center', va='bottom', fontweight='bold')
# Adjust layout
plt.tight_layout()
# Save chart
timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
chart_filename = f"rag_vs_direct_latency_comparison_{timestamp}.png"
# Ensure results directory exists
results_dir = Path(__file__).parent / "results"
results_dir.mkdir(exist_ok=True)
chart_path = results_dir / chart_filename
plt.savefig(chart_path, dpi=300, bbox_inches='tight',
facecolor='white', edgecolor='none')
plt.close()
print(f"π RAG vs Direct latency comparison charts saved to: {chart_path}")
return str(chart_path)
except Exception as e:
print(f"β Chart generation failed: {e}")
return ""
def print_comparison_summary(self, rag_stats: Dict, direct_stats: Dict):
"""Print formatted comparison summary to console"""
print(f"\nπ === RAG vs DIRECT LATENCY COMPARISON SUMMARY ===")
# Overall comparison
rag_overall = rag_stats['overall_results']
direct_overall = direct_stats['overall_results']
print(f"\nπ Overall Performance:")
print(f" RAG System:")
print(f" β’ Average Latency: {rag_overall['average_latency']:.2f}s")
print(f" β’ Success Rate: {rag_overall['successful_queries']}/{rag_overall['total_queries']} ({rag_overall['successful_queries']/rag_overall['total_queries']*100:.1f}%)")
print(f" β’ 60s Target Compliance: {rag_overall['target_compliance']*100:.1f}%")
print(f" Direct LLM System:")
print(f" β’ Average Latency: {direct_overall['average_latency']:.2f}s")
print(f" β’ Success Rate: {direct_overall['successful_queries']}/{direct_overall['total_queries']} ({direct_overall['success_rate']*100:.1f}%)")
print(f" β’ 60s Target Compliance: {direct_overall['target_compliance']*100:.1f}%")
# Performance winner
if direct_overall['average_latency'] < rag_overall['average_latency']:
latency_winner = "Direct LLM"
latency_improvement = rag_overall['average_latency'] - direct_overall['average_latency']
else:
latency_winner = "RAG"
latency_improvement = direct_overall['average_latency'] - rag_overall['average_latency']
print(f"\nπ Performance Winner:")
print(f" β’ Faster System: {latency_winner}")
print(f" β’ Performance Improvement: {latency_improvement:.2f}s ({latency_improvement/max(rag_overall['average_latency'], direct_overall['average_latency'])*100:.1f}%)")
# Category breakdown
print(f"\nπ Category Breakdown:")
categories = ['diagnosis', 'treatment', 'mixed']
for category in categories:
rag_cat = rag_stats['category_results'].get(category, {})
direct_cat = direct_stats['category_results'].get(category, {})
if rag_cat.get('query_count', 0) > 0 and direct_cat.get('query_count', 0) > 0:
rag_latency = rag_cat.get('average_latency', 0)
direct_latency = direct_cat.get('average_latency', 0)
winner = "Direct" if direct_latency < rag_latency else "RAG"
difference = abs(rag_latency - direct_latency)
print(f" {category.capitalize()}:")
print(f" β’ RAG: {rag_latency:.2f}s")
print(f" β’ Direct: {direct_latency:.2f}s")
print(f" β’ Winner: {winner} (faster by {difference:.2f}s)")
# Independent execution interface
if __name__ == "__main__":
"""Independent chart generation interface"""
print("π OnCall.ai RAG vs Direct Latency Comparison Chart Generator")
# Initialize chart generator
chart_gen = RAGvsDirectLatencyChartGenerator()
try:
# Find latest statistics files
rag_file, direct_file = chart_gen.find_latest_statistics_files()
# Load statistics
rag_stats, direct_stats = chart_gen.load_statistics(rag_file, direct_file)
# Generate comparison charts
print(f"π Generating RAG vs Direct comparison charts...")
chart_path = chart_gen.generate_comparison_charts(rag_stats, direct_stats)
# Print comparison summary
chart_gen.print_comparison_summary(rag_stats, direct_stats)
print(f"\nβ
RAG vs Direct latency comparison complete!")
print(f"π Charts saved to: {chart_path}")
print(f"π‘ Charts optimized for research presentations and publications")
except FileNotFoundError as e:
print(f"β {e}")
print("π‘ Please ensure both evaluators have been run:")
print(" python latency_evaluator.py # for RAG statistics")
print(" python direct_llm_evaluator.py # for Direct statistics")
except Exception as e:
print(f"β Chart generation failed: {e}")
|