test: fix cost harness review issues

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>

tests/integration/_support/upstream.py

@@ -1,19 +1,25 @@
 from __future__ import annotations
 
 import argparse
-import json
-import os
-import struct
-import zlib
 from collections import deque
 from collections.abc import Mapping
+import json
 from dataclasses import dataclass, field
+import os
 from pathlib import Path
 from queue import SimpleQueue
+import struct
 from typing import Final, cast
+import zlib
 
 import httpx
 import uvicorn
+from pydantic import BaseModel, JsonValue, TypeAdapter, ValidationError
+from starlette.applications import Starlette
+from starlette.requests import Request
+from starlette.responses import JSONResponse, Response
+from starlette.routing import Route
+
 from _fake_openai_endpoint_server import chat_completions, completions, embeddings, health, moderations
 from integration.cost_calculation.cost_tracking_case import (
     EventStreamResponse,
@@ -21,11 +27,6 @@ from integration.cost_calculation.cost_tracking_case import (
     SseResponse,
     StoredResponse,
 )
-from pydantic import BaseModel, JsonValue, TypeAdapter, ValidationError
-from starlette.applications import Starlette
-from starlette.requests import Request
-from starlette.responses import JSONResponse, Response
-from starlette.routing import Route
 
 JSON_OBJECT: Final = TypeAdapter(dict[str, JsonValue])
 CASES_FILE: Final = Path(__file__).resolve().parents[1] / "cost_calculation" / "cost_tracking_cases.json"

tests/integration/cost_calculation/conftest.py

@@ -9,11 +9,12 @@ from typing import Final
 
 from cryptography.hazmat.primitives import serialization
 from cryptography.hazmat.primitives.asymmetric import rsa
+from pydantic import BaseModel, ConfigDict
+
 from integration._support.client import JSON_OBJECT, Scenario, eventually, object_value, string_value
 from integration._support.database import read_rows
 from integration._support.upstream import delete_scenario, register_scenario
 from integration.cost_calculation.cost_tracking_case import CostTrackingTestCase
-from pydantic import BaseModel, ConfigDict
 
 
 class CostBreakdown(BaseModel):

tests/integration/cost_calculation/cost_tracking_cases.json

@@ -21817,8 +21817,10 @@
       "response": {
         "content_type": "application/json",
         "body": {
+          "id": "resp_$REQUEST_ID",
           "object": "response",
           "status": "completed",
+          "created_at": 1700000000,
           "model": "gpt-5.6",
           "output": [
             {

tests/integration/cost_calculation/test_cost_tracking.py

@@ -6,6 +6,7 @@ from hashlib import sha256
 from typing import Final, cast
 
 import pytest
+
 from integration._support.client import JSON_OBJECT, Gateway
 from integration.cost_calculation.conftest import (
     approx_equal,
@@ -105,15 +106,34 @@ def test_case_bills_expected_cost(gateway: Gateway, case: CostTrackingTestCase)
         assert breakdown.output_cost is not None and approx_equal(breakdown.output_cost, expected.output_cost), (
             f"{case.name}: output_cost {breakdown.output_cost} != expected {expected.output_cost}"
         )
-        for field, header_name, expected_component in (
-            ("cache_read_cost", "x-litellm-response-cost-cache-read", expected.cache_read_cost),
-            ("cache_creation_cost", "x-litellm-response-cost-cache-creation", expected.cache_creation_cost),
-            ("reasoning_cost", "x-litellm-response-cost-reasoning", expected.reasoning_cost),
-            ("tool_usage_cost", "x-litellm-response-cost-tool-usage", expected.tool_usage_cost),
+        for field, header_name, actual_component, expected_component in (
+            (
+                "cache_read_cost",
+                "x-litellm-response-cost-cache-read",
+                breakdown.cache_read_cost,
+                expected.cache_read_cost,
+            ),
+            (
+                "cache_creation_cost",
+                "x-litellm-response-cost-cache-creation",
+                breakdown.cache_creation_cost,
+                expected.cache_creation_cost,
+            ),
+            (
+                "reasoning_cost",
+                "x-litellm-response-cost-reasoning",
+                breakdown.reasoning_cost,
+                expected.reasoning_cost,
+            ),
+            (
+                "tool_usage_cost",
+                "x-litellm-response-cost-tool-usage",
+                breakdown.tool_usage_cost,
+                expected.tool_usage_cost,
+            ),
         ):
             if expected_component is None:
                 continue
-            actual_component: Final = getattr(breakdown, field)
             assert actual_component is not None and approx_equal(actual_component, expected_component), (
                 f"{case.name}: {field} {actual_component} != expected {expected_component}"
             )