refactor: optimize token calculation in context command#292
Merged
Conversation
- Use API-reported values instead of tokenizer estimation where possible - Remove reasoning category (absorbed into assistant as residual) - Batch tokenizer calls for efficiency (4 calls instead of N) - System: derived from first assistant input - first user tokens - User: tokenized (batched) - Tools: tokenized (batched, needed for pruning anyway) - Assistant: calculated as residual (total - system - user - tools) - Add detailed documentation explaining the calculation strategy
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
/dcp contextto minimize tokenizer estimationChanges
Token calculation strategy:
firstAssistant.input + cache.read - tokenizer(firstUser)tokenizer(all user messages)- batched into single calltokenizer(inputs + outputs) - pruned- batched into 2 callstotal - system - user - tools(residual)Why assistant is the residual:
If reasoning tokens persist in context (model-dependent behavior varies by provider), they semantically belong with "Assistant" since reasoning IS assistant-generated content.
Removed:
reasoningcategory from breakdown (absorbed into assistant residual)Documentation
Added detailed comment block at top of
context.tsexplaining the full calculation strategy.