fix(langchain): sanitize SummarizationMiddleware history to prevent token overflow
#34557
Conversation
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
github-actions
bot
added
langchain
sridhar852002
changed the title
github-actions
bot
added
fix
mdrxy
changed the title
SummarizationMiddleware history to prevent token overflow
github-actions
bot
added
fix
Member
|
Thanks for the contribution @sridhar852002! Your analysis of the root cause was spot on - the issue is that
Implementing the fix in #34607, which is a simpler change (just updating the import and using the existing utility) while achieving the same result. Closing this in favor of that approach, but thanks again for identifying the issue and proposing a solution! |
mdrxy
added a commit
that referenced
this pull request
Jan 7, 2026
#34607) Fixes #34517 Supersedes #34557, #34570 Fixes token inflation in `SummarizationMiddleware` that caused context window overflow during summarization. **Root cause:** When formatting messages for the summary prompt, `str(messages)` was implicitly called, which includes all Pydantic metadata fields (`usage_metadata`, `response_metadata`, `additional_kwargs`, etc.). This caused the stringified representation to use ~2.5x more tokens than `count_tokens_approximately` estimates. **Problem:** - Summarization triggers at 85% of context window based on `count_tokens_approximately` - But `str(messages)` in the prompt uses 2.5x more tokens - Results in `ContextLengthExceeded` **Fix:** Use `get_buffer_string()` to format messages, which produces compact output: ``` Human: What's the weather? AI: Let me check...[tool_calls] Tool: 72°F and sunny ``` Instead of verbose Pydantic repr: ```python [HumanMessage(content='What's the weather?', additional_kwargs={}, response_metadata={}), ...] ```
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.
Description
Fixes #34517.
Context
Currently,
SummarizationMiddlewaresends the full message objects to the LLM when generating a summary. This includes all metadata (liketoken_usage,logprobs, etc.), which wastes a significant amount of tokens. This triggersContextLengthExceedederrors even when the actual text content is small.Changes
_format_clean_historyto strip metadata before sending history to the summarization model.Role: Contentstring format to save tokens.Checklist
make lintandmake format