Fix dropped Inkling reasoning at stream end

[ispobock]

Motivation

With stream_reasoning=False, the Inkling reasoning detector buffers the thinking block into _pending_reasoning and only flushes it when a control/end token closes the block. When generation is cut mid-block (e.g. max_tokens truncates a <|content_thinking|> block), the stream ends with no end token, so the buffered trace is dropped entirely — the whole reasoning disappears from the streamed response.

The non-streaming detect_and_parse path already flushes _pending_reasoning at the end, but InklingDetector had no finish() override, so at stream end parse_stream_end() fell through to the base finish() (which only handles _force_nonempty_content/_in_reasoning, neither used by this detector) and emitted nothing.

Modifications

• Override InklingDetector.finish() to emit any buffered _pending_reasoning (when mid-reasoning under stream_reasoning=False) and reset the detector's block state, mirroring the non-streaming flush.
• Add a regression test that streams a thinking block truncated before its end token and asserts finish() surfaces the trailing reasoning (fails on the pre-fix code, passes after).

Verified on a live server that BFCL live tool-calling accuracy is unchanged (tool_parse_errors=0), i.e. the added stream-end flush does not perturb the tool-call path.

---

CI States

Latest PR Test (Base): 🚫 Run #29753147981
Latest PR Test (Extra): ❌ Run #29753146335

[ispobock]

/rerun-test test/registered/unit/parser/test_reasoning_parser.py

[github-actions]

Results for /rerun-test test/registered/unit/parser/test_reasoning_parser.py:

🚀 ubuntu-latest (1 test): ✅ View workflow run

cd test/ && python3 registered/unit/parser/test_reasoning_parser.py

[gemini-code-assist]

Code Review

This pull request adds a finish method to the reasoning parser to flush buffered reasoning when stream_reasoning=False and the stream ends before a control token closes the block, and includes a unit test to verify this behavior. The reviewer points out that the base class BaseReasoningFormatDetector suffers from the same issue and suggests updating its finish method to ensure consistent behavior across all reasoning detectors.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.