File size: 13,882 Bytes
d1b209f | 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 | #!/usr/bin/env python3
"""
Generate comprehensive statistics report for SciVisAgentBench.
Processes all CSV files and creates updated statistics.
"""
import pandas as pd
import os
from collections import defaultdict
def load_csv_files(sheets_dir):
"""Load all CSV files from the sheets directory."""
csv_files = {}
# Define which files to include (excluding old chatvis_bench and main)
files_to_include = [
'SciVisAgentBench_Statistics - bioimage_data.csv',
'SciVisAgentBench_Statistics - molecular_vis.csv',
'SciVisAgentBench_Statistics - sci_volume_data.csv',
'SciVisAgentBench_Statistics - topology.csv',
'SciVisAgentBench_Statistics - paraview.csv' # New file
]
for filename in files_to_include:
filepath = os.path.join(sheets_dir, filename)
if os.path.exists(filepath):
df = pd.read_csv(filepath)
# Extract key from filename
key = filename.replace('SciVisAgentBench_Statistics - ', '').replace('.csv', '')
csv_files[key] = df
print(f"Loaded {filename}: {len(df)} rows")
return csv_files
def count_individual_tags(series, separator=';'):
"""Count individual tags in a series of semicolon-separated values."""
counts = defaultdict(int)
for value in series.dropna():
if pd.isna(value) or str(value).strip() == '':
continue
# Support both ; and , as separators
value_str = str(value).replace(',', ';')
tags = [tag.strip() for tag in value_str.split(separator)]
for tag in tags:
if tag:
counts[tag] += 1
return dict(sorted(counts.items(), key=lambda x: -x[1]))
def count_combinations(series):
"""Count exact combinations as they appear."""
counts = defaultdict(int)
for value in series.dropna():
if pd.isna(value) or str(value).strip() == '':
continue
counts[str(value).strip()] += 1
return dict(sorted(counts.items(), key=lambda x: -x[1]))
def generate_report(csv_files):
"""Generate comprehensive statistics report."""
# Combine all dataframes
all_data = pd.concat(csv_files.values(), ignore_index=True)
# Filter only Task and Workflow levels (exclude Operation level)
cases_only = all_data[all_data['Task Level 1: Complexity Level'].isin(['Task', 'Workflow'])]
print(f"\nTotal rows in all files: {len(all_data)}")
print(f"Total cases (Task + Workflow): {len(cases_only)}")
# Generate statistics
stats = {}
# 1. Total Cases Count
complexity_counts = all_data['Task Level 1: Complexity Level'].value_counts()
stats['total_cases'] = len(cases_only)
stats['total_tasks'] = complexity_counts.get('Task', 0)
stats['total_workflows'] = complexity_counts.get('Workflow', 0)
stats['total_operations'] = complexity_counts.get('Operation', 0)
# File breakdown
stats['file_breakdown'] = {}
for name, df in csv_files.items():
# Calculate total operation count
cases_df = df[df['Task Level 1: Complexity Level'].isin(['Task', 'Workflow'])]
# Convert Operation Count to numeric, treating 'N/A' as NaN
operation_counts = pd.to_numeric(cases_df['Operation Count'], errors='coerce')
total_operations = operation_counts.sum() if len(operation_counts) > 0 else 0
file_stats = {
'total_rows': len(df),
'operations': len(df[df['Task Level 1: Complexity Level'] == 'Operation']),
'tasks': len(df[df['Task Level 1: Complexity Level'] == 'Task']),
'workflows': len(df[df['Task Level 1: Complexity Level'] == 'Workflow']),
'cases': len(cases_df),
'total_operations': int(total_operations)
}
stats['file_breakdown'][name] = file_stats
# 2. Application Domain Statistics (cases only)
stats['applications_individual'] = count_individual_tags(cases_only['Application'])
stats['applications_combinations'] = count_combinations(cases_only['Application'])
# 3. Data Type Statistics (cases only)
stats['data_types_individual'] = count_individual_tags(cases_only['Data'])
stats['data_types_combinations'] = count_combinations(cases_only['Data'])
# 4. Visualization Operations Statistics (cases only)
stats['operations_all'] = count_individual_tags(cases_only['Task Level 2: Visualization Operations'], separator=';')
# Operations by complexity level
tasks_only = cases_only[cases_only['Task Level 1: Complexity Level'] == 'Task']
workflows_only = cases_only[cases_only['Task Level 1: Complexity Level'] == 'Workflow']
stats['operations_task_level'] = count_individual_tags(tasks_only['Task Level 2: Visualization Operations'], separator=';')
stats['operations_workflow_level'] = count_individual_tags(workflows_only['Task Level 2: Visualization Operations'], separator=';')
# Also count operation combinations
stats['operations_combinations'] = count_combinations(cases_only['Task Level 2: Visualization Operations'])
return stats, cases_only, all_data
def write_markdown_report(stats, output_path):
"""Write statistics report to markdown file."""
with open(output_path, 'w') as f:
f.write("# SciVisAgentBench - Comprehensive Statistics Report\n")
f.write(f"*Generated from {len(stats['file_breakdown'])} CSV files in the benchmark*\n")
f.write("---\n")
# 1. Total Cases Count
f.write("## 1. Total Cases Count\n")
f.write("**Important**: Cases = Tasks + Workflows only (Operation-level entries are NOT counted as cases)\n\n")
f.write("### Overall Summary\n")
f.write(f"- **Total Tasks**: {stats['total_tasks']}\n")
f.write(f"- **Total Workflows**: {stats['total_workflows']}\n")
f.write(f"- **Total Cases**: **{stats['total_cases']}** (Tasks + Workflows)\n")
if stats['total_operations'] > 0:
f.write(f"- **Total Operations**: **{stats['total_operations']}** (Operation-level entries)\n")
f.write("\n")
f.write("### Breakdown by File\n")
f.write("| File | Tasks | Workflows | Cases (Task+Workflow) | Total Operations |\n")
f.write("|------|-------|-----------|----------------------|-----------------|\n")
total_tasks = 0
total_workflows = 0
total_cases = 0
total_operations_sum = 0
for name, file_stats in stats['file_breakdown'].items():
f.write(f"| {name} | {file_stats['tasks']} | {file_stats['workflows']} | "
f"**{file_stats['cases']}** | {file_stats['total_operations']} |\n")
total_tasks += file_stats['tasks']
total_workflows += file_stats['workflows']
total_cases += file_stats['cases']
total_operations_sum += file_stats['total_operations']
f.write(f"| **TOTAL** | **{total_tasks}** | **{total_workflows}** | **{total_cases}** | **{total_operations_sum}** |\n")
f.write("\n")
# 2. Application Domain Statistics
f.write("## 2. Application Domain Statistics\n")
f.write("**Note**: These statistics include ONLY cases (Tasks + Workflows). Operation-level entries are excluded.\n\n")
f.write("### Individual Application Counts\n")
f.write("*(Counts individual tags, so multi-tagged entries contribute to multiple categories)*\n\n")
f.write("| Application | Count |\n")
f.write("|-------------|-------|\n")
for app, count in stats['applications_individual'].items():
f.write(f"| {app} | {count} |\n")
total_app_tags = sum(stats['applications_individual'].values())
f.write(f"\n**Total individual application tags**: {total_app_tags}\n\n")
f.write("### Application Combinations\n")
f.write("*(Shows exact combinations as they appear in the data)*\n\n")
f.write("| Application Combination | Count |\n")
f.write("|------------------------|-------|\n")
for combo, count in stats['applications_combinations'].items():
f.write(f"| {combo} | {count} |\n")
f.write("\n")
# 3. Data Type Statistics
f.write("## 3. Data Type Statistics\n")
f.write("**Note**: These statistics include ONLY cases (Tasks + Workflows). Operation-level entries are excluded.\n\n")
f.write("### Individual Data Type Counts\n")
f.write("*(Counts individual tags, so multi-tagged entries contribute to multiple categories)*\n\n")
f.write("| Data Type | Count |\n")
f.write("|-----------|-------|\n")
for dtype, count in stats['data_types_individual'].items():
f.write(f"| {dtype} | {count} |\n")
total_dtype_tags = sum(stats['data_types_individual'].values())
f.write(f"\n**Total individual data type tags**: {total_dtype_tags}\n\n")
f.write("### Data Type Combinations\n")
f.write("*(Shows exact combinations as they appear in the data)*\n\n")
f.write("| Data Type Combination | Count |\n")
f.write("|-----------------------|-------|\n")
for combo, count in stats['data_types_combinations'].items():
f.write(f"| {combo} | {count} |\n")
f.write("\n")
# 4. Complexity Level Statistics
f.write("## 4. Task Level 1: Complexity Level Statistics\n\n")
f.write("### Overall Distribution\n")
f.write("| Complexity Level | Entry Count | Counted as Case? |\n")
f.write("|------------------|-------------|------------------|\n")
if stats['total_operations'] > 0:
f.write(f"| Operation | {stats['total_operations']} | ❌ NO |\n")
f.write(f"| Task | {stats['total_tasks']} | ✅ YES |\n")
f.write(f"| Workflow | {stats['total_workflows']} | ✅ YES |\n")
f.write(f"| **Total Cases** | **{stats['total_cases']}** | **(Tasks + Workflows)** |\n")
f.write("\n")
# 5. Visualization Operations Statistics
f.write("## 5. Task Level 2: Visualization Operations Statistics\n")
f.write("**Note**: These statistics include ONLY cases (Tasks + Workflows). Operation-level entries are excluded.\n\n")
f.write("### All Visualization Operations (Sorted by Frequency)\n")
f.write("| Rank | Visualization Operation | Total Count |\n")
f.write("|------|------------------------|-------------|\n")
for i, (op, count) in enumerate(stats['operations_all'].items(), 1):
f.write(f"| {i} | {op} | {count} |\n")
total_op_tags = sum(stats['operations_all'].values())
f.write(f"\n**Total visualization operation tags**: {total_op_tags}\n\n")
f.write("### Top 10 Most Common Visualization Operations\n")
f.write("| Rank | Operation | Count |\n")
f.write("|------|-----------|-------|\n")
for i, (op, count) in enumerate(list(stats['operations_all'].items())[:10], 1):
f.write(f"| {i} | {op} | {count} |\n")
f.write("\n")
f.write("### Visualization Operations by Complexity Level (Cases Only)\n\n")
f.write("#### Task Level (Top 10)\n")
f.write("| Rank | Operation | Count |\n")
f.write("|------|-----------|-------|\n")
for i, (op, count) in enumerate(list(stats['operations_task_level'].items())[:10], 1):
f.write(f"| {i} | {op} | {count} |\n")
f.write("\n")
f.write("#### Workflow Level (Top 10)\n")
f.write("| Rank | Operation | Count |\n")
f.write("|------|-----------|-------|\n")
for i, (op, count) in enumerate(list(stats['operations_workflow_level'].items())[:10], 1):
f.write(f"| {i} | {op} | {count} |\n")
f.write("\n")
# 6. Summary Statistics
f.write("## 6. Summary Statistics\n")
f.write(f"- **Total number of CSV files analyzed**: {len(stats['file_breakdown'])}\n")
f.write(f"- **Total Cases (Tasks + Workflows)**: **{stats['total_cases']}**\n")
f.write(f"- **Unique application domains**: {len(stats['applications_individual'])}\n")
f.write(f"- **Unique data types**: {len(stats['data_types_individual'])}\n")
f.write(f"- **Unique visualization operations**: {len(stats['operations_all'])}\n\n")
f.write("### File Contributions\n")
f.write("| File | Cases Contributed | Percentage |\n")
f.write("|------|-------------------|------------|\n")
for name, file_stats in sorted(stats['file_breakdown'].items(),
key=lambda x: -x[1]['cases']):
percentage = (file_stats['cases'] / stats['total_cases'] * 100) if stats['total_cases'] > 0 else 0
f.write(f"| {name} | {file_stats['cases']} | {percentage:.1f}% |\n")
f.write("\n")
def main():
# Paths
sheets_dir = '/Users/kuangshiai/Documents/ND-VIS/Code/SciVisAgentBench-tasks/statistics/sheets'
output_path = '/Users/kuangshiai/Documents/ND-VIS/Code/SciVisAgentBench-tasks/statistics/statistics_report.md'
# Load CSV files
print("Loading CSV files...")
csv_files = load_csv_files(sheets_dir)
# Generate statistics
print("\nGenerating statistics...")
stats, cases_only, all_data = generate_report(csv_files)
# Write markdown report
print(f"\nWriting report to {output_path}...")
write_markdown_report(stats, output_path)
print("\n✅ Report generation complete!")
print(f" Total cases: {stats['total_cases']}")
print(f" Total files: {len(csv_files)}")
if __name__ == '__main__':
main()
|