Organize Open-SWE-Traces data prep project

This commit is contained in:
Codex
2026-06-24 22:34:04 +08:00
commit f06e573b04
27 changed files with 3264 additions and 0 deletions

View File

@@ -0,0 +1,32 @@
#!/usr/bin/env python3
import json
from pathlib import Path
import pyarrow.parquet as pq
BASE = Path("data/Open-SWE-Traces/data")
CONFIGS = [
"minimax_m25_openhands_trajectories",
"minimax_m25_sweagent_trajectories",
"qwen35_openhands_trajectories",
"qwen35_sweagent_trajectories",
]
def main() -> int:
for cfg in CONFIGS:
file = sorted((BASE / cfg).glob("*.parquet"))[0]
row = pq.read_table(file).slice(0, 1).to_pylist()[0]
print("CFG", cfg)
print("cols", list(row.keys()))
trajectory = row["trajectory"]
print("len", len(trajectory))
for index, message in enumerate(trajectory[:10]):
print("MSG", index, json.dumps(message, ensure_ascii=False)[:1600])
print("---")
return 0
if __name__ == "__main__":
raise SystemExit(main())