File size: 5,008 Bytes
e25024e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
45113e6
e25024e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
#!/usr/bin/env python3
"""Aggregate community contributions into paper-ready tables.

Usage:
    python scripts/aggregate_contributions.py [--dir community_results] [--format latex|csv|json]

Reads all contribution JSON files from the specified directory, aggregates
them by model and method, and outputs summary tables suitable for inclusion
in the paper.
"""

from __future__ import annotations

import argparse
import json
import sys
from pathlib import Path

# Add project root to path
sys.path.insert(0, str(Path(__file__).resolve().parent.parent))

from obliteratus.community import (
    aggregate_results,
    generate_latex_table,
    load_contributions,
)


def main():
    parser = argparse.ArgumentParser(
        description="Aggregate community contributions into paper tables."
    )
    parser.add_argument(
        "--dir",
        default="community_results",
        help="Directory containing contribution JSON files (default: community_results)",
    )
    parser.add_argument(
        "--format",
        choices=["latex", "csv", "json", "summary"],
        default="summary",
        help="Output format (default: summary)",
    )
    parser.add_argument(
        "--metric",
        default="refusal_rate",
        help="Metric to display in tables (default: refusal_rate)",
    )
    parser.add_argument(
        "--methods",
        nargs="*",
        help="Methods to include (default: all)",
    )
    parser.add_argument(
        "--min-runs",
        type=int,
        default=1,
        help="Minimum runs per (model, method) to include (default: 1)",
    )
    args = parser.parse_args()

    # Load all contributions
    records = load_contributions(args.dir)
    if not records:
        print(f"No contributions found in {args.dir}/", file=sys.stderr)
        sys.exit(1)

    print(f"Loaded {len(records)} contribution(s) from {args.dir}/", file=sys.stderr)

    # Aggregate
    aggregated = aggregate_results(records)

    # Filter by minimum runs
    if args.min_runs > 1:
        for model in list(aggregated.keys()):
            for method in list(aggregated[model].keys()):
                if aggregated[model][method]["n_runs"] < args.min_runs:
                    del aggregated[model][method]
            if not aggregated[model]:
                del aggregated[model]

    if not aggregated:
        print("No results meet the minimum run threshold.", file=sys.stderr)
        sys.exit(1)

    # Output
    if args.format == "summary":
        _print_summary(aggregated, args.metric)
    elif args.format == "latex":
        print(generate_latex_table(aggregated, methods=args.methods, metric=args.metric))
    elif args.format == "json":
        print(json.dumps(aggregated, indent=2))
    elif args.format == "csv":
        _print_csv(aggregated, args.metric)


def _print_summary(aggregated: dict, metric: str):
    """Print a human-readable summary of aggregated results."""
    total_runs = sum(
        data["n_runs"]
        for model_data in aggregated.values()
        for data in model_data.values()
    )
    n_models = len(aggregated)
    n_methods = len(set(
        method
        for model_data in aggregated.values()
        for method in model_data
    ))

    print(f"\n{'=' * 70}")
    print("Community Contribution Summary")
    print(f"{'=' * 70}")
    print(f"  Total runs:    {total_runs}")
    print(f"  Models:        {n_models}")
    print(f"  Methods:       {n_methods}")
    print()

    for model in sorted(aggregated.keys()):
        model_data = aggregated[model]
        short = model.split("/")[-1] if "/" in model else model
        print(f"  {short}:")
        for method in sorted(model_data.keys()):
            data = model_data[method]
            n = data["n_runs"]
            if metric in data:
                stats = data[metric]
                mean = stats["mean"]
                std = stats["std"]
                if std > 0 and n > 1:
                    print(f"    {method:20s}  {metric}={mean:.2f} ± {std:.2f}  (n={n})")
                else:
                    print(f"    {method:20s}  {metric}={mean:.2f}  (n={n})")
            else:
                print(f"    {method:20s}  (no {metric} data, n={n})")
        print()

    print(f"{'=' * 70}")
    print(f"To generate LaTeX: python {sys.argv[0]} --format latex")
    print(f"To generate CSV:   python {sys.argv[0]} --format csv")


def _print_csv(aggregated: dict, metric: str):
    """Print results as CSV."""
    print("model,method,n_runs,mean,std,min,max")
    for model in sorted(aggregated.keys()):
        for method in sorted(aggregated[model].keys()):
            data = aggregated[model][method]
            n = data["n_runs"]
            if metric in data:
                stats = data[metric]
                print(
                    f"{model},{method},{n},"
                    f"{stats['mean']:.4f},{stats['std']:.4f},"
                    f"{stats['min']:.4f},{stats['max']:.4f}"
                )


if __name__ == "__main__":
    main()