Add injectable token counter; conservative default estimate - #11
Merged
Conversation
Context-window trimming relied on a fixed ~4 chars/token heuristic that under-counts code/JSON/non-ASCII, risking over-window prompts. - Conversation.tokenCounter: optional synchronous hook that overrides the heuristic — plug in an exact per-provider tokenizer for precise trimming. - Default heuristic is now conservative: charsPerToken 4.0 -> 3.5 plus a per-message overhead (tokensPerMessageOverhead, default 4). Adds regression tests for the override and the overhead. 90 tests pass. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
Caution The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased. |
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Roadmap P1.4 — real tokenizer (#5). A multi-provider kit shouldn't hardcode one tokenizer (each provider tokenizes differently), so this adds a hook plus a safer default.
Conversation.tokenCounter— an optional synchronous counter that overrides the built-in heuristic used for context-window trimming. Apps can plug in an exact per-provider tokenizer (or cached count-tokens results).charsPerToken4.0 → 3.5 and a per-message framing overhead (tokensPerMessageOverhead, default 4), so trimming errs toward staying under the real window for code/JSON/non-ASCII histories (which the old heuristic under-counted).Regression tests for the override and the overhead. 90 tests pass. Behavior change is limited to trimming being slightly more aggressive by default; both knobs are public and tunable.
🤖 Generated with Claude Code