Konstantinos Thermos 28ffc814c0 fix(llm): retry OpenAI import with max_completion_tokens ...

Newer OpenAI models (gpt-5.x) reject max_tokens with a 400 and demand
max_completion_tokens, so AI booking import failed and no reservations
were created. Send max_tokens first (understood by the classic chat API
and every local server), and only on a 400 naming max_completion_tokens
resend the whole request with that parameter instead — mirroring the
existing json_schema->json_object fallback beside it. Detection is by
API response, not a model-name allowlist.

Fixes #1760

2026-08-05 15:13:17 +02:00

..

e2e

feat(server): count only started trips as visited in the atlas

2026-08-03 20:21:18 +02:00

fixtures

v3.4.0 (#1540)

2026-07-18 22:13:52 +02:00

helpers

chore: move notifications backend to proper DI

2026-08-02 11:36:06 +02:00

integration

feat(server): count only started trips as visited in the atlas

2026-08-03 20:21:18 +02:00

unit

fix(llm): retry OpenAI import with max_completion_tokens

2026-08-05 15:13:17 +02:00

websocket

chore: move auth backend to proper DI

2026-08-01 17:39:59 +02:00

setup.console-noise.ts

fix(server): silence expected-error noise in test stderr

2026-07-28 22:31:46 +02:00

setup.ts

fix(server): close the notification-send teardown race in tests

2026-07-28 22:31:40 +02:00