Skip to content

Freeze Python tool timeout in RL requests #2

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
dajiaohuang wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Freeze Python tool timeout in RL requests #2

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
082146e
Freeze Python tool timeout in RL handoff
dajiaohuang Sep 25, 2026
11b5ee6
test: assert frozen RL tool timeout
dajiaohuang Sep 26, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions src/simple_scibuddy/coevolve/loop.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -186,6 +186,7 @@ def run(config, cfg):
"max_output_tokens": hcfg["max_tokens"],
"context_tokens": hcfg["context_tokens"],
"rollout_timeout_seconds": hcfg["seconds"],
"tool_seconds": cfg["tool_seconds"],
"tool_response_tokens": hcfg["tool_response_tokens"],
"tool_history_tokens": hcfg["tool_history_tokens"],
"train_index": str(root / "train.jsonl"),
Expand Down
3 changes: 2 additions & 1 deletion src/simple_scibuddy/coevolve/worker.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,7 +39,8 @@ def plan_round(cfg, directory, steps=50):
no_eval=False, no_checkpoint=False, eval_samples=1, eval_temperature=0.0)
mapping = {"context_tokens": "context_tokens", "max_output_tokens": "max_tokens",
"max_model_calls": "actions", "rollout_timeout_seconds": "seconds",
"tool_response_tokens": "tool_response_tokens", "tool_history_tokens": "tool_history_tokens"}
"tool_seconds": "tool_seconds", "tool_response_tokens": "tool_response_tokens",
"tool_history_tokens": "tool_history_tokens"}
for source, destination in mapping.items():
value = contract[source]
if not isinstance(value, int) or isinstance(value, bool) or value < 1:
Expand Down
5 changes: 3 additions & 2 deletions tests/test_explicit_dataset.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,7 +75,7 @@ def test_rl_handoff_accepts_only_declared_train_test_ids(tmp_path, monkeypatch):
contract = dict(thinking_enabled=False, execution_network='disabled', temperature=1.0, top_p=1.0,
top_k=-1, entrypoint='run(task, api)', max_model_calls=9, max_tool_calls=9,
context_tokens=24576, max_output_tokens=4096, rollout_timeout_seconds=900,
tool_response_tokens=2048, tool_history_tokens=8192)
tool_seconds=47, tool_response_tokens=2048, tool_history_tokens=8192)
for split, key in [('train', 'train_index'), ('test', 'eval_index')]:
path = tmp_path / f'{split}.jsonl'
path.write_text(''.join(json.dumps(r) + '\n' for r in rows if r['split'] == split))
Expand All @@ -85,7 +85,8 @@ def test_rl_handoff_accepts_only_declared_train_test_ids(tmp_path, monkeypatch):
write_json(tmp_path / 'request.json', request)
monkeypatch.setattr(protocol, 'read_request', lambda _: request)
monkeypatch.setattr(worker, 'validate_payloads', lambda *args: None)
worker.plan_round({'dataset': str(tmp_path / 'dataset')}, tmp_path, 30)
_, planned = worker.plan_round({'dataset': str(tmp_path / 'dataset'), 'tool_seconds': 999}, tmp_path, 30)
assert planned['tool_seconds'] == contract['tool_seconds'] == 47
path = tmp_path / 'train.jsonl'
path.write_text(path.read_text().replace('train-0', 'val-0'))
with pytest.raises(ValueError, match='differs from frozen'):
Expand Down