Add Google AI model to NL Git assistant endpoint
What does this MR do and why?
It is now possible to receive GoogleAI response by sending model=vertexai
param to /ai/llm/git_command
API endpoint. We need both at the moment to smoothly migrate Gitlab CLI to GoogleAI response. A feature flag for this change is not introduced because the change is non-breaking. The whole feature is already behind a feature flag (ai_git_command_ff
).
Related issue: https://gitlab.com/gitlab-org/gitlab/-/issues/414275+
Edited by Igor Drozdov