test(vertex_ai): count Google token fetches for RAG Engine searches through the proxy

This commit is contained in:
mateo-berri 2026-10-29 14:13:03 -07:00
parent 5ccf2ed6bf
commit 7e768e06cb

View file

@ -4,6 +4,7 @@ from typing import Final
from urllib.parse import urlsplit
import httpx
import pytest
from integration._support.client import Gateway
from integration._support.vertex import service_account_json
from integration._support.wire import Reply, Request, wire_server
@ -145,8 +146,21 @@ def test_vector_store_search_forwards_all_search_fields_on_both_routes(gateway:
]
def test_vertex_search_api_vector_store_reuses_one_google_token_across_searches(gateway: Gateway) -> None:
store_id: Final = f"ds_search_{uuid.uuid4().hex}"
@pytest.mark.parametrize(
("api_type", "search_target", "search_reply"),
(
(
"search_api",
"/store:search",
{"results": [{"id": "doc-1", "document": {"id": "doc-1", "derivedStructData": {"title": "hit"}}}]},
),
("rag_api", "/store:retrieveContexts", {"contexts": {"contexts": []}}),
),
)
def test_vertex_vector_store_reuses_one_google_token_across_searches(
gateway: Gateway, api_type: str, search_target: str, search_reply: dict[str, JsonValue]
) -> None:
store_id: Final = f"vs_vertex_{uuid.uuid4().hex}"
def respond(request: Request) -> Reply:
assert request.method == "POST"
@ -154,13 +168,9 @@ def test_vertex_search_api_vector_store_reuses_one_google_token_across_searches(
return Reply(
body=json.dumps({"access_token": _VERTEX_TOKEN, "token_type": "Bearer", "expires_in": 3600}).encode()
)
assert request.target == "/search:search", request.target
assert request.target == search_target, request.target
assert request.headers["authorization"] == f"Bearer {_VERTEX_TOKEN}"
return Reply(
body=json.dumps(
{"results": [{"id": "doc-1", "document": {"id": "doc-1", "derivedStructData": {"title": "hit"}}}]}
).encode()
)
return Reply(body=json.dumps(search_reply).encode())
with wire_server(respond) as wire, gateway.scenario() as scenario:
registered: Final = gateway.request(
@ -168,9 +178,9 @@ def test_vertex_search_api_vector_store_reuses_one_google_token_across_searches(
"/vector_store/new",
{
"vector_store_id": store_id,
"custom_llm_provider": "vertex_ai/search_api",
"custom_llm_provider": f"vertex_ai/{api_type}",
"litellm_params": {
"api_base": f"{wire.url}/search",
"api_base": f"{wire.url}/store",
"vertex_project": _VERTEX_PROJECT,
"vertex_location": "global",
"vertex_credentials": service_account_json(_VERTEX_PROJECT, wire.url),
@ -190,5 +200,5 @@ def test_vertex_search_api_vector_store_reuses_one_google_token_across_searches(
assert [request.target for request in wire.drain()] == [
"/_oauth/token",
*("/search:search",) * _VERTEX_SEARCH_COUNT,
*(search_target,) * _VERTEX_SEARCH_COUNT,
]