7b139e28d4aee15881250d85659c0d1b41eed9c0
49%

Ran 17 Dec 2025 01:30AM UTC

Jobs 0

Files 0

Run time –

Badge

Embed ▾

pending completion

cancel

Build # 7b139e28d4aee15881250d85659c0d1b41eed9c0

Build Type

push

github

Committed by

web-flow

Commit Message

feat(vertex): Add Google Search grounding support for Gemini models (#284)

* feat(vertex): Add Google Search grounding support for Gemini models

Vertex AI requires camelCase for grounding tools (googleSearch) while
the Google AI API uses snake_case (google_search). This adds:

- google_grounding option to Vertex AI provider schema
- Transformation layer in Gemini formatter to convert snake_case keys
  to camelCase before sending to Vertex AI
- Unit tests for the grounding tool transformation

* fix(vertex): Extract google_grounding from nested provider_options

Google.encode_body expects google_grounding at request.options[:google_grounding]
but it was nested under provider_options and never reaching the grounding
tools builder. This extracts and hoists it to the top level.

* fix(vertex): Remove incorrect camelCase transformation for grounding tools

Vertex AI REST API uses snake_case (google_search) not camelCase
(googleSearch). The previous transformation was breaking grounding.

Also removes redundant opts hoisting since Options.process already
hoists provider_options to top level.

Run Details

Source Files on build 7b139e28d4aee15881250d85659c0d1b41eed9c0

Detailed source file information is not available for this build.

agentjido / req_llm / 7b139e28d4aee15881250d85659c0d1b41eed9c0
49%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Source Files on build 7b139e28d4aee15881250d85659c0d1b41eed9c0

agentjido / req_llm / 7b139e28d4aee15881250d85659c0d1b41eed9c0 49%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Source Files on build 7b139e28d4aee15881250d85659c0d1b41eed9c0

agentjido / req_llm / 7b139e28d4aee15881250d85659c0d1b41eed9c0
49%

README BADGES
x