feat(compare): 比价记录 LLM token 成本落库与展示(按当时价冻结) (#133)

- comparison_record 加 llm_cost_yuan(元/float)+ llm_price_snapshot(JSON)两列
- _backfill_llm_calls 回填时按 app_config 当时单价逐模型算成本、冻结成本+快照到记录
- app_config 新增 llm_token_price 配置(per_model + default 兜底,运营在系统配置页可改)
- services/llm_cost.py:compute_llm_cost 纯函数(按 model 分桶、error/无 usage 跳过、
  脏价格当 unpriced 不抛异常以免连累 token 回填)+ get_llm_prices reader
- admin schema 暴露成本:列表项带 llm_cost_yuan,详情另带价格快照
- tests/test_llm_cost.py(10 测试);scripts/seed_mock_llm_cost.py(mock seeder)

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

---------

Co-authored-by: guke <guke@autohome.com.cn>
Reviewed-on: #133

app/api/v1/compare_record.py

@@ -27,6 +27,7 @@ from app.schemas.compare_record import (
     ComparisonRecordOut,
     ComparisonRecordPage,
 )
+from app.services.llm_cost import compute_llm_cost, get_llm_prices
 from app.services.pricebot_llm_calls import fetch_llm_calls
 
 logger = logging.getLogger("shagua.compare_record")
@@ -81,6 +82,8 @@ def _backfill_llm_calls(record_id: int, trace_id: str) -> None:
         # error 的调用 usage 可能为 None,or {} 兜底)
         rec.input_tokens = sum((c.get("usage") or {}).get("prompt_tokens") or 0 for c in calls)
         rec.output_tokens = sum((c.get("usage") or {}).get("completion_tokens") or 0 for c in calls)
+        # 本次比价 LLM 成本(元)+ 当时单价快照:按 app_config 现价逐模型算好冻结(services/llm_cost.py)。
+        rec.llm_cost_yuan, rec.llm_price_snapshot = compute_llm_cost(calls, get_llm_prices(db))
         db.commit()
         logger.info(
             "backfill llm_calls trace=%s n=%d in_tok=%d out_tok=%d",