Yufeng He 8f95ca9d98 fix: filter Gemini thinking parts from user-facing message chain (#7196) ...

Gemini 3 models return thinking parts (part.thought=True) alongside the
actual response text.  _process_content_parts was including these thinking
parts in the message chain sent to the user, effectively leaking internal
reasoning into the output.  On platforms that split long messages (e.g.
aiocqhttp with realtime segmenting), this caused duplicate or triple
replies since the thinking text often mirrors the actual response.

The streaming path already handled this correctly via chunk.text which
skips thinking parts, but the non-streaming path and the final-chunk
processing in streaming both went through _process_content_parts.

Also switch the Gemini 3 model name matching from an exhaustive list to
prefix matching (gemini-3- / gemini-3.) so new variants like gemini-3.1
get proper thinkingLevel config without code changes.

Fixes #7183

2026-04-03 16:35:25 +08:00

..

api

feat: add plugin load&unload hook (#5331)

2026-02-23 23:13:41 +08:00

builtin_stars

fix(pipeline): skip waking on empty messages (#6893)

2026-03-28 20:50:42 +08:00

cli

chore: update version to 4.22.2 in multiple files

2026-03-29 00:32:31 +08:00

core

fix: filter Gemini thinking parts from user-facing message chain (#7196)

2026-04-03 16:35:25 +08:00

dashboard

fix: 改进知识库的初始化错误处理 (#7243)

2026-04-02 16:16:22 +08:00

utils

fix: stabilize packaged runtime pip/ssl behavior and mac font fallback (#5007)

2026-02-10 16:42:43 +09:00

__init__.py

…