feat: yaml_patcher and ProfileApplyRequest support metric_weights and doc_weights

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This commit is contained in:
2026-06-18 17:02:21 +08:00
parent 480f6d66ea
commit ce0d2291b0
4 changed files with 349 additions and 2 deletions

View File

@@ -2,13 +2,18 @@
from __future__ import annotations
import time
from fastapi import APIRouter, HTTPException
from openai import OpenAI
from webapp.models import (
CreateProfileRequest,
LLMProfile,
ProfileApplyRequest,
ProfileApplyResponse,
ProfileProbeRequest,
ProfileTestResponse,
)
from webapp.services.profile_manager import profile_manager
from webapp.services.yaml_patcher import apply_profiles_to_scenario
@@ -16,6 +21,43 @@ from webapp.services.yaml_patcher import apply_profiles_to_scenario
router = APIRouter(prefix="/api/llm-profiles", tags=["llm-profiles"])
def _do_connectivity_test(
model: str,
base_url: str,
api_key: str,
timeout_seconds: int,
) -> ProfileTestResponse:
"""Send a minimal chat completion request and return the test result."""
client = OpenAI(
api_key=api_key,
base_url=base_url.rstrip("/"),
timeout=float(timeout_seconds),
)
t0 = time.monotonic()
try:
client.chat.completions.create(
model=model,
messages=[{"role": "user", "content": "hi"}],
max_tokens=1,
)
latency_ms = int((time.monotonic() - t0) * 1000)
return ProfileTestResponse(ok=True, message="连接成功", latency_ms=latency_ms)
except Exception as exc: # noqa: BLE001
latency_ms = int((time.monotonic() - t0) * 1000)
return ProfileTestResponse(ok=False, message=str(exc), latency_ms=latency_ms)
@router.post("/probe", response_model=ProfileTestResponse, tags=["llm-profiles"])
def probe_connectivity(request: ProfileProbeRequest) -> ProfileTestResponse:
"""Test LLM connectivity with inline credentials (no saved profile required)."""
return _do_connectivity_test(
model=request.model,
base_url=request.base_url,
api_key=request.api_key,
timeout_seconds=request.timeout_seconds,
)
@router.get("", response_model=dict)
def list_profiles() -> dict:
"""Return all saved LLM profiles."""
@@ -59,6 +101,20 @@ def delete_profile(profile_id: str) -> dict:
return {"deleted": True}
@router.post("/{profile_id}/test", response_model=ProfileTestResponse)
def test_profile(profile_id: str) -> ProfileTestResponse:
"""Test LLM connectivity for a saved profile."""
profile = profile_manager.get(profile_id)
if profile is None:
raise HTTPException(status_code=404, detail=f"Profile not found: {profile_id}")
return _do_connectivity_test(
model=profile.model,
base_url=profile.base_url,
api_key=profile.api_key,
timeout_seconds=profile.timeout_seconds,
)
@router.post("/apply", response_model=ProfileApplyResponse)
def apply_profiles(request: ProfileApplyRequest) -> ProfileApplyResponse:
"""Patch selected LLM profiles into the target scenario YAML file."""
@@ -89,6 +145,8 @@ def apply_profiles(request: ProfileApplyRequest) -> ProfileApplyResponse:
judge_profile=role_profiles["judge"],
answer_profile=role_profiles["answer"],
dataset_profile=role_profiles["dataset"],
metric_weights=request.metric_weights,
doc_weights=request.doc_weights,
)
return ProfileApplyResponse(
scenario_path=request.scenario_path,