scriptoriumadmin/litellm
1
0
Fork 0
mirror of https://github.com/BerriAI/litellm.git synced 2026-10-06 02:48:13 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 103

test(e2e/claude_code): scope the GPT cell opt-in rationale to the cron VM

Browse source
This commit is contained in:
mateo-berri 2026-07-15 16:27:58 -07:00
parent 710a88eba7
commit 90f7807830
1 changed files with 9 additions and 9 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

18
tests/e2e/claude_code/_gpt_cells.py
View file

@ -16,15 +16,15 @@ cover "OpenAI plus the big three clouds":
carries only the open-weight
gpt-oss MaaS models
Live GPT cells are opt-in via `COMPAT_GPT_CELLS=1`. The external PR
gate and the daily cron VM must be provisioned with the GPT-route
credentials (`OPENAI_API_KEY`, `AZURE_OPENAI_API_BASE` +
`AZURE_OPENAI_API_KEY`, and Bedrock Mantle model access) before these
cells can pass, so until the flag is set each live cell skips and its
matrix cell stays `not_tested` — landing this suite change cannot flip
the existing gate red. The `vertex_ai_gpt` column ignores the flag:
its cells report a static `not_applicable` and never touch the
network.
Live GPT cells are opt-in via `COMPAT_GPT_CELLS=1`. The cron VM that
runs the scheduled suite and publishes the matrix must be provisioned
with the GPT-route credentials (`OPENAI_API_KEY` with available
quota, `AZURE_OPENAI_API_BASE` + `AZURE_OPENAI_API_KEY` with gpt-5.6
deployments, and Bedrock Mantle model access) before these cells can
pass, so until the flag is set each live cell skips and its matrix
cell publishes as `not_tested` instead of a credential-shaped red.
The `vertex_ai_gpt` column ignores the flag: its cells report a
static `not_applicable` and never touch the network.
"""
from __future__ import annotations