WeKnora/internal
wizardchen 9e08ab7302 fix(chat): map MaxTokens to MaxCompletionTokens for GPT-5/o-series
OpenAI's GPT-5 series and o-series reasoning models (o1/o3/o4-mini) no
longer accept `max_tokens` and reject non-default sampling params
(`temperature`, `top_p`, `frequency_penalty`, `presence_penalty`).
Azure OpenAI propagates the same constraint, returning HTTP 400 with
"this model is not supported MaxTokens, please use MaxCompletionTokens".

Lower the compatibility shim into `RemoteAPIChat.BuildChatCompletionRequest`
(the single OpenAI-protocol egress) so every internal caller keeps using
`MaxTokens` uniformly:

- Add `provider.IsOpenAIReasoningOrGPT5Model` with precise prefix matching
  (covers `gpt-5*`, `o1`/`o1-*`, `o3`/`o3-*`, `o4`/`o4-*`; rejects
  `olympus-1`, `openai-*`, `o3xtra`, etc.).
- When the provider is `openai`/`azure_openai` and the model matches, map
  `MaxTokens` to `MaxCompletionTokens` (explicit `MaxCompletionTokens`
  wins) and skip the unsupported sampling fields; `omitempty` keeps them
  off the wire.
- Behavior is unchanged for gpt-4o, gpt-4, and all other providers.

Add unit tests for the matcher and for the request-building path
(Azure gpt-5.2, OpenAI gpt-5/o1-mini/o3/o4-mini, plus negative cases).

Fixes #1283
2026-05-13 00:40:35 +08:00
..
agent fix(agent): exclude wiki-only KBs from quick-answer (RAG) mode 2026-05-12 16:27:28 +08:00
application feat(observability): expand Langfuse spans across chat pipeline 2026-05-12 21:16:37 +08:00
assets feat(assets): add ASR test audio file and embed it in the application 2026-04-02 21:27:27 +08:00
common feat: Implement deadlock retry mechanism for chunk creation 2026-04-22 21:17:21 +08:00
config feat(agent): human-in-the-loop approval for MCP tool calls (#1173) 2026-05-10 22:57:12 +08:00
container refactor(agent): rebuild multi-turn history from DB, drop llmcontext layer 2026-05-12 12:26:24 +08:00
database fix: update environment configuration for debugging and logging, enhance SQLite FTS5 table management, and improve API base URL handling 2026-04-11 18:51:16 +08:00
datasource fix(feishu): tolerate partial wiki node listing failures 2026-05-12 17:40:59 +08:00
errors feat: Add agent configuration and cleanup scripts for database migrations 2025-11-05 23:18:44 +08:00
event feat(agent): human-in-the-loop approval for MCP tool calls (#1173) 2026-05-10 22:57:12 +08:00
handler feat(observability): log end-to-end TTFB on both ends of chat stream 2026-05-12 21:16:37 +08:00
im fix(agent): exclude wiki-only KBs from quick-answer (RAG) mode 2026-05-12 16:27:28 +08:00
infrastructure refactor(web-search): drop dead api_key path in SearXNG provider 2026-05-11 16:53:47 +08:00
logger feat(observability): integrate Langfuse for LLM token tracking and tracing 2026-04-24 10:29:19 +08:00
mcp feat(mcp): implement reconnection logic for MCP tool calls and tool listing 2026-03-31 11:57:15 +08:00
middleware fix(middleware): mask camelCase secret fields in request logs 2026-05-12 19:27:29 +08:00
models fix(chat): map MaxTokens to MaxCompletionTokens for GPT-5/o-series 2026-05-13 00:40:35 +08:00
router feat(agent): human-in-the-loop approval for MCP tool calls (#1173) 2026-05-10 22:57:12 +08:00
runtime init commit 2025-08-05 15:08:07 +08:00
sandbox feat: optimize security and deployment of agent skills 2026-02-04 20:08:49 +08:00
searchutil feat(prompt): enhance question rewriting and intent classification in prompt templates 2026-04-16 23:15:02 +08:00
stream feat(redis): add REDIS_USERNAME support for Redis ACL 2026-02-04 19:38:40 +08:00
tracing feat(observability): extend Langfuse tracing across asynq pipeline 2026-04-24 13:16:47 +08:00
types feat(agent): support dedicated model for query understanding step 2026-05-12 20:27:20 +08:00
utils refactor(security_test): standardize struct field formatting in SSRF whitelist tests 2026-05-11 16:53:47 +08:00