Commit Graph
6 Commits
Author SHA1 Message Date
Lingao MengandXubin Ren 0e617c32cd fix(shell): kill subprocess on CancelledError to prevent orphan processes
When an agent task is cancelled (e.g. via /stop), the ExecTool was only
handling TimeoutError but not CancelledError. This left the child process
running as an orphan. Now CancelledError also triggers process.kill() and
waitpid cleanup before re-raising.
2026-04-06 13:47:09 +08:00
Lingao MengandXubin Ren 519911456a test(provider): fix incorrect assertion in reasoning_content sanitize test
The test test_openai_compat_strips_message_level_reasoning_fields was
added in fbedf7a and incorrectly asserted that reasoning_content and
extra_content should be stripped from messages. This contradicts the
intent of b5302b6 which explicitly added these fields to _ALLOWED_MSG_KEYS
to preserve them through sanitization.

Rename the test and fix assertions to match the original design intent:
reasoning_content and extra_content at message level should be preserved,
and extra_content inside tool_calls should also be preserved.

Signed-off-by: Lingao Meng <menglingao@xiaomi.com>
2026-04-04 20:08:44 +08:00
Lingao MengandXubin Ren 3f8eafc89a fix(provider): restore reasoning_content and extra_content in message sanitization
reasoning_content and extra_content were accidentally dropped from
_ALLOWED_MSG_KEYS.

Also fix session/manager.py to include reasoning_content when building
LLM messages from session history, so the field is not lost across
turns.

Without this fix, providers such as Kimi, emit reasoning_content in
assistant messages will have it stripped on the next request, breaking
multi-turn thinking mode.

Fixes: https://github.com/HKUDS/nanobot/issues/2777
Signed-off-by: Lingao Meng <menglingao@xiaomi.com>
2026-04-04 20:08:44 +08:00
Lingao MengandXubin Ren a05f83da89 test(providers): cover reasoning_content extraction in OpenAI compat provider
Add regression tests for the non-streaming (_parse dict branch) and
streaming (_parse_chunks dict and SDK-object branches) paths that extract
reasoning_content, ensuring the field is populated when present and None
when absent.

Signed-off-by: Lingao Meng <menglingao@xiaomi.com>
2026-04-04 02:09:57 +08:00
Lingao MengandXubin Ren 210643ed68 feat(provider): support reasoning_content in OpenAI compat provider
Extract reasoning_content from both non-streaming and streaming responses
in OpenAICompatProvider. Accumulate chunks during streaming and merge into
LLMResponse, enabling reasoning chain display for models like MiMo and DeepSeek-R1.

Signed-off-by: Lingao Meng <menglingao@xiaomi.com>
2026-04-04 02:09:57 +08:00
Lingao Meng cf6c979339 feat(provider): add Xiaomi MiMo LLM support
Register xiaomi_mimo as an OpenAI-compatible provider with its API base URL,
add xiaomi_mimo to the provider config schema, and document it in README.

Signed-off-by: Lingao Meng <menglingao@xiaomi.com>
2026-04-03 14:42:57 +08:00