litellm/tests/unit/llms/vertex_ai
devin-ai-integration[bot] 3fa688223d
fix(vertex_ai): translate /v1/responses batch rows through the Responses-to-Chat bridge (#43042)
* fix(vertex_ai): translate /v1/responses batch rows through the Responses-to-Chat bridge

Vertex batch uploads treated every non-embeddings JSONL row as a chat
completions body, so a /v1/responses row lost its input and reached GCS
as a blank text part. Route detection now recognizes /v1/responses rows
and bridges them to chat through the same Responses-to-Chat bridge the
real-time path uses. That bridge call moves out of the Bedrock files
transformation into a shared helper both providers call, forwarding the
record's fields as sent, like real time, instead of validating them
against the SDK TypedDicts whose required keys clients omit.

* chore(batches): type the Vertex responses test helper and drop the quoted input cast

* fix(batches): translate developer messages to system on Vertex and Bedrock batch rows like real time

---------

Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com>
2026-09-24 16:15:58 -07:00
..
agent_engine test: migrate legacy provider tests to tests/unit (wave 2, phase 13) 2026-09-20 14:18:49 +00:00
context_caching test: restore pagination limit tests and realtime query param regression test 2026-09-20 14:53:26 +00:00
files fix(vertex_ai): translate /v1/responses batch rows through the Responses-to-Chat bridge (#43042) 2026-09-24 16:15:58 -07:00
gemini_embeddings test: migrate legacy provider tests to tests/unit (wave 2, phase 13) 2026-09-20 14:18:49 +00:00
image_edit test: migrate legacy provider tests to tests/unit (wave 2, phase 13) 2026-09-20 14:18:49 +00:00
interactions test: migrate legacy provider tests to tests/unit (wave 2, phase 13) 2026-09-20 14:18:49 +00:00
multimodal_embeddings test: migrate legacy provider tests to tests/unit (wave 2, phase 13) 2026-09-20 14:18:49 +00:00
realtime test: restore pagination limit tests and realtime query param regression test 2026-09-20 14:53:26 +00:00
text_to_speech fix(logging): pass provider response headers to callbacks on every endpoint (#42824) 2026-09-24 13:01:12 -07:00
vertex_ai_partner_models test: add __init__.py to every tests/unit directory this migration touches 2026-09-20 11:52:09 +00:00
vertex_gemma_models test: migrate phase 14 wave 2 provider tests to tests/unit 2026-09-20 11:01:40 +00:00
videos test: migrate phase 14 wave 2 provider tests to tests/unit 2026-09-20 11:01:40 +00:00
__init__.py test: migrate legacy provider tests to tests/unit (wave 2, phase 13) 2026-09-20 14:18:49 +00:00