File size: 1,893 Bytes
430d13c 76eacb1 430d13c 5f867e3 430d13c 5f867e3 430d13c df6fa5a 430d13c | 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 | """Write the arena submission body from the live /schema answer and the task list."""
import argparse
import json
import sys
from pathlib import Path
from urllib.request import urlopen
sys.path.insert(0, str(Path(__file__).resolve().parents[1]))
from workbench.tasks import TASK_IDS # noqa: E402
LIMITS = {"split": "train", "reset_wall_s": 180, "rollout_wall_s": 1800, "verifier_wall_s": 120,
"tool_wall_s": 120, "tool_calls_total": 64, "completion_tokens": 32768,
"context_tokens": 32768, "memory_gib": 4, "cpu_floor_vcpus": 1, "workspace_gib": 2}
def main():
p = argparse.ArgumentParser()
p.add_argument("--url", default="http://127.0.0.1:18000")
p.add_argument("--image", required=True)
p.add_argument("--id", default="analyst-workbench-v1")
p.add_argument("--out", default="submission.json")
p.add_argument("--tasks", default="", help="comma-separated task ids (default: every family at level 1)")
a = p.parse_args()
tasks = a.tasks.split(",") if a.tasks else TASK_IDS
schema = json.load(urlopen(a.url + "/schema"))
body = {
"submission_id": a.id,
"name": "Analyst Workbench",
"image": a.image,
"schema": {"action": schema["action"], "observation": schema["observation"]},
"tasks": [dict(LIMITS, task_id=t, **({"verifier_wall_s": 300} if t.startswith("swe-repo-") else {}))
for t in tasks],
"example_actions": [{"tool": "ls"}, {"tool": "submit", "answer": ""}],
"finish_action": {"tool": "submit", "answer": ""},
"dataset": "ajinkyamulay/analyst-workbench",
"source": "https://huggingface.co/datasets/ajinkyamulay/analyst-workbench/tree/main/source",
}
Path(a.out).write_text(json.dumps(body, indent=2) + "\n")
print(f"wrote {a.out}: {len(body['tasks'])} tasks, image {a.image}")
if __name__ == "__main__":
main()
|