File size: 3,885 Bytes
638c806 | 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 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 | from __future__ import annotations
from fastapi.testclient import TestClient
from promptboost.agentic import AgenticPrepassError, boost_agentic
from promptboost.main import app
from promptboost.schemas import AgenticTaskDraft
class FakeLLM:
def __init__(
self,
draft: AgenticTaskDraft | None = None,
raw_output: str | None = None,
error: Exception | None = None,
model: str = "fake-agentic-model",
) -> None:
self.draft = draft
self.raw_output = raw_output
self.error = error
self.model = model
def create_task_draft(self, raw_prompt: str, target_harness: str, target_model: str):
if self.error:
raise self.error
assert self.draft is not None
return self.draft, self.raw_output or self.draft.model_dump_json(), self.model
def _safe_draft() -> AgenticTaskDraft:
return AgenticTaskDraft(
task_type="coding_patch",
goal="Fix the API bug and add tests without unrelated refactors.",
inputs=["API bug", "tests"],
deliverables=["Fix the API bug", "add tests"],
constraints=["without unrelated refactors"],
success_criteria=["tests"],
open_questions=[],
assumptions=[],
risk_flags=[],
confidence=0.86,
)
def test_agentic_prepass_adds_validated_draft(temp_db) -> None:
result = boost_agentic(
"Fix the API bug and add tests without unrelated refactors.",
"codex",
"gpt",
llm_client=FakeLLM(_safe_draft()),
)
assert result.agentic is not None
assert result.agentic.status == "accepted"
assert result.no_new_facts_passed is True
assert "## Agentic Task Draft" in result.boosted_prompt
assert result.agentic.task_draft is not None
assert result.agentic.task_draft.confidence == 0.86
def test_agentic_prepass_falls_back_on_llm_error(temp_db) -> None:
result = boost_agentic(
"Summarize the provided source text without adding facts.",
"codex",
"gpt",
llm_client=FakeLLM(error=AgenticPrepassError("timeout")),
)
assert result.agentic is not None
assert result.agentic.status == "fallback"
assert "## Agentic Task Draft" not in result.boosted_prompt
assert any("Agentic PromptBoost fallback" in warning for warning in result.warnings)
def test_agentic_prepass_discards_new_fact_draft(temp_db) -> None:
draft = AgenticTaskDraft(
task_type="document_summary",
goal="Summarize the provided notes about Kubernetes latency in Berlin.",
inputs=["provided notes"],
deliverables=["Summarize the provided notes"],
constraints=[],
success_criteria=[],
open_questions=[],
assumptions=[],
risk_flags=["Kubernetes latency in Berlin"],
confidence=0.5,
)
result = boost_agentic(
"Summarize the provided notes.",
"codex",
"gpt",
llm_client=FakeLLM(draft),
)
assert result.agentic is not None
assert result.agentic.status == "fallback"
assert result.agentic.reason == "agentic_no_new_facts_failed"
assert "## Agentic Task Draft" not in result.boosted_prompt
def test_agentic_route_falls_back_when_api_key_missing(temp_db, monkeypatch) -> None:
monkeypatch.delenv("PROMPTBOOST_LLM_API_KEY", raising=False)
client = TestClient(app)
response = client.post(
"/boost/agentic",
json={
"raw_prompt": "Fix the API bug and add tests without unrelated refactors.",
"target_harness": "codex",
"target_model": "gpt",
},
)
assert response.status_code == 200
body = response.json()
assert body["result"]["agentic"]["status"] == "fallback"
assert "PROMPTBOOST_LLM_API_KEY" in body["result"]["agentic"]["reason"]
assert body["run"]["id"] == 1
|