open-swe/tests/test_agent_subagent_models.py
Adam Moussa 1c93e5c62b
feat: switch model providers to AWS Bedrock (Claude) and Fireworks (non-Claude)
Migrate off direct provider APIs: AWS Bedrock for Anthropic/Claude via the
cross-region inference profile us.anthropic.claude-opus-4-8, Fireworks AI for
all non-Claude models. Drop OpenAI (gpt-5.5) and Google (gemini-3.5-flash)
entirely. DEFAULT_MODEL_ID is now Bedrock Claude; all Fireworks models stay
freely selectable for the agent and reviewer graphs and via team/profile
defaults.

- pyproject: add langchain-aws (ChatBedrockConverse + boto3)
- options.py: Bedrock Claude entry + default; remove openai/google entries
- model.py: bedrock_converse provider_model_kwargs (effort -> thinking budget),
  region pin in make_model, bedrock<->fireworks fallback pairing, AWS_REGION/
  FIREWORKS_API_KEY local-dev validation
- server.py: provider-aware fallback kwargs build
- sanitize_thinking_blocks: also sanitize ChatBedrockConverse thinking blocks
- model_fallback: treat transient botocore ClientError codes as fallback-worthy
- eval_jobs: repoint hardcoded eval model id to Bedrock Claude
- tests: repoint dropped model ids; drop obsolete google test module
2026-06-29 14:57:14 -04:00

157 lines
5.7 KiB
Python

from unittest.mock import AsyncMock, MagicMock, patch
import pytest
from langgraph.graph.state import RunnableConfig
from agent.server import get_agent
class _DummyAgent:
def with_config(self, config: RunnableConfig) -> "_DummyAgent":
self.config = config
return self
@pytest.mark.asyncio
async def test_agent_uses_profile_subagent_model_override() -> None:
config: RunnableConfig = {
"configurable": {
"__is_for_execution__": True,
"thread_id": "thread-123",
"github_login": "octocat",
},
"metadata": {},
}
main_model = MagicMock(name="main_model")
subagent_model = MagicMock(name="subagent_model")
captured: dict[str, object] = {}
def fake_create_deep_agent(**kwargs: object) -> _DummyAgent:
captured.update(kwargs)
return _DummyAgent()
with (
patch(
"agent.server.resolve_github_token",
new_callable=AsyncMock,
return_value=("ghp", None),
),
patch("agent.server.resolve_triggering_user_identity", return_value=None),
patch(
"agent.server.ensure_sandbox_for_thread",
new_callable=AsyncMock,
return_value=MagicMock(),
),
patch(
"agent.server.aresolve_sandbox_work_dir",
new_callable=AsyncMock,
return_value="/workspace",
),
patch(
"agent.server.get_team_default_model_pair",
new_callable=AsyncMock,
return_value=(
("bedrock_converse:us.anthropic.claude-opus-4-8", "medium"),
("fireworks:accounts/fireworks/models/deepseek-v4-pro", "low"),
),
),
patch(
"agent.server.load_profile",
new_callable=AsyncMock,
return_value={
"default_model": "bedrock_converse:us.anthropic.claude-opus-4-8",
"reasoning_effort": "high",
"default_subagent_model": "fireworks:accounts/fireworks/models/deepseek-v4-pro",
"subagent_reasoning_effort": "xhigh",
},
),
patch("agent.server.fallback_model_id_for", return_value=None),
patch("agent.server.make_model", side_effect=[main_model, subagent_model]) as make_model,
patch("agent.server.construct_system_prompt", return_value="prompt"),
patch("agent.server.create_deep_agent", side_effect=fake_create_deep_agent),
):
await get_agent(config)
assert captured["model"] is main_model
subagents = captured["subagents"]
assert isinstance(subagents, list)
assert subagents[0]["name"] == "general-purpose"
assert subagents[0]["model"] is subagent_model
main_call = make_model.call_args_list[0]
assert main_call.args == ("bedrock_converse:us.anthropic.claude-opus-4-8",)
assert main_call.kwargs["additional_model_request_fields"] == {
"thinking": {"type": "enabled", "budget_tokens": 16000}
}
subagent_call = make_model.call_args_list[1]
assert subagent_call.args == ("fireworks:accounts/fireworks/models/deepseek-v4-pro",)
assert subagent_call.kwargs["model_kwargs"] == {"reasoning_effort": "xhigh"}
@pytest.mark.asyncio
async def test_agent_subagent_inherits_profile_model_override_without_explicit_pair() -> None:
config: RunnableConfig = {
"configurable": {
"__is_for_execution__": True,
"thread_id": "thread-123",
"github_login": "octocat",
},
"metadata": {},
}
main_model = MagicMock(name="main_model")
subagent_model = MagicMock(name="subagent_model")
captured: dict[str, object] = {}
def fake_create_deep_agent(**kwargs: object) -> _DummyAgent:
captured.update(kwargs)
return _DummyAgent()
with (
patch(
"agent.server.resolve_github_token",
new_callable=AsyncMock,
return_value=("ghp", None),
),
patch("agent.server.resolve_triggering_user_identity", return_value=None),
patch(
"agent.server.ensure_sandbox_for_thread",
new_callable=AsyncMock,
return_value=MagicMock(),
),
patch(
"agent.server.aresolve_sandbox_work_dir",
new_callable=AsyncMock,
return_value="/workspace",
),
patch(
"agent.server.get_team_default_model_pair",
new_callable=AsyncMock,
return_value=(
("bedrock_converse:us.anthropic.claude-opus-4-8", "medium"),
("fireworks:accounts/fireworks/models/deepseek-v4-pro", "low"),
),
),
patch(
"agent.server.load_profile",
new_callable=AsyncMock,
return_value={
"default_model": "bedrock_converse:us.anthropic.claude-opus-4-8",
"reasoning_effort": "high",
},
),
patch("agent.server.fallback_model_id_for", return_value=None),
patch("agent.server.make_model", side_effect=[main_model, subagent_model]) as make_model,
patch("agent.server.construct_system_prompt", return_value="prompt"),
patch("agent.server.create_deep_agent", side_effect=fake_create_deep_agent),
):
await get_agent(config)
subagents = captured["subagents"]
assert isinstance(subagents, list)
assert subagents[0]["model"] is subagent_model
assert make_model.call_args_list[0].args == ("bedrock_converse:us.anthropic.claude-opus-4-8",)
assert make_model.call_args_list[1].args == ("bedrock_converse:us.anthropic.claude-opus-4-8",)
assert make_model.call_args_list[1].kwargs["additional_model_request_fields"] == {
"thinking": {"type": "enabled", "budget_tokens": 16000}
}