fix(reflect): cap structured output retry budget - #2433
Merged
nicoloboschi merged 1 commit intoJun 29, 2026
Merged
Conversation
koriyoshi2041
force-pushed
the
rios/cap-reflect-structured-retries
branch
from
June 27, 2026 07:23
ccef070 to
689633f
Compare
nicoloboschi
approved these changes
Jun 29, 2026
r266-tech
added a commit
to r266-tech/hindsight
that referenced
this pull request
Jul 1, 2026
…raction vectorize-io#2433 capped the structured retry budget but the structured second pass never received an output-token budget, so on reasoning/preamble models the provider default is exhausted before JSON is emitted (finish_reason=length, empty content) and structured reflect degrades to None. Thread the reflect max_tokens through _generate_structured_output (and _process_done_tool) as max_completion_tokens, mirroring the plain reflect calls. Fixes vectorize-io#2431.
nicoloboschi
pushed a commit
that referenced
this pull request
Jul 20, 2026
… extraction call (#2431) (#2486) * fix(reflect): thread max_completion_tokens into structured-output extraction #2433 capped the structured retry budget but the structured second pass never received an output-token budget, so on reasoning/preamble models the provider default is exhausted before JSON is emitted (finish_reason=length, empty content) and structured reflect degrades to None. Thread the reflect max_tokens through _generate_structured_output (and _process_done_tool) as max_completion_tokens, mirroring the plain reflect calls. Fixes #2431. * test(reflect): cover structured-output max_completion_tokens threading
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
reflect_structuredextraction retry budget so provider-specific empty structured-output responses do not consume the full reflect timeoutContext
This is motivated by #2431, where plain reflect can succeed while
response_schematriggers repeated empty provider responses withfinish_reason=length.Tests
uv run pytest tests/test_reflect_agent.py::TestReflectStructuredOutput::test_structured_output_uses_short_retry_budget -quv run ruff check hindsight_api/engine/reflect/agent.py tests/test_reflect_agent.pyuv run ruff format --check hindsight_api/engine/reflect/agent.py tests/test_reflect_agent.pygit diff --checkNote: full
uv run pytest tests/test_reflect_agent.py -qreached 50 passing tests, then hit local environment setup errors unrelated to this patch: missingHINDSIGHT_API_LLM_API_KEYfor real-LLM tests and a local SOCKS proxy requiringsocksioduring Hugging Face model initialization.