File size: 2,848 Bytes
c95bb11 | 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 | #!/usr/bin/env python3
"""research_diff.py - Compare decision registers across parallel trial runners.
Usage:
python3 research_diff.py TRIAL_DIR_1 TRIAL_DIR_2 [TRIAL_DIR_3 ...]
python3 research_diff.py run_a run_b run_c
Outputs a clean markdown comparison showing where models agreed and where choices split.
"""
import argparse
import csv
import sys
from pathlib import Path
REQUIRED_STAGES = ["question", "scope", "data", "digest", "draft", "mint"]
def load_decisions(trial_path):
p = Path(trial_path) / "decisions.csv"
if not p.exists():
sys.exit(f"Error: {p} does not exist")
with p.open(encoding="utf-8", newline="") as fh:
reader = csv.DictReader(fh)
rows = {r["decision_id"].strip().upper(): r for r in reader if r.get("decision_id")}
return rows
def main():
parser = argparse.ArgumentParser(description="Compare decisions across trial directories.")
parser.add_argument("trials", nargs="+", help="Paths to trial directories to compare")
parser.add_argument("--markdown", action="store_true", default=True, help="Output in markdown format")
args = parser.parse_args()
if len(args.trials) < 2:
sys.exit("Error: Specify at least two trial directories to diff.")
trial_dirs = [Path(t) for t in args.trials]
trial_data = {t.name: load_decisions(t) for t in trial_dirs}
all_dids = []
for d in trial_data.values():
for did in d.keys():
if did not in all_dids and did:
all_dids.append(did)
names = [t.name for t in trial_dirs]
print(f"# Cross-Runner Decision Diff")
print(f"Comparing {len(names)} runners: {', '.join(names)}\n")
for did in all_dids:
rows = [trial_data[name].get(did, {}) for name in names]
question_text = next((r.get("question") for r in rows if r.get("question")), did)
stage = next((r.get("stage") for r in rows if r.get("stage")), "")
chosen_vals = [r.get("chosen", "").strip() for r in rows]
has_content = any(chosen_vals)
if not has_content:
continue
all_identical = len(set(chosen_vals)) == 1
print(f"## {did} ({stage}: {question_text})")
status = "CONSENSUS" if all_identical else "DIVERGENCE"
print(f"**Status:** {status}\n")
for name, r in zip(names, rows):
by = r.get("decided_by", "unspecified")
chosen = r.get("chosen", "[empty]")
alts = r.get("alternatives", "")
rationale = r.get("rationale", "")
print(f"- **{name}** (`{by}`):")
print(f" - *Chosen:* {chosen}")
if alts:
print(f" - *Alternatives:* {alts}")
if rationale:
print(f" - *Rationale:* {rationale}")
print("\n---\n")
if __name__ == "__main__":
main()
|