Conversation
Fixes: #16577 From 14.0 up until now, our correction-search focused on the minimum "correction distance" to a possible text _correction_, then performed prediction _on top of that_. This PR changes the search to integrate word-frequency weighting from models within the search, allowing us to optimize for the most likely predictions. There is no additional performace cost impact of note here - the probability associated with a correction's best prediction is already accessible O(1) with operations we've already been doing. It is likely that this will help further mitigate issues with transposition handling noted in the user tests for #16450. Build-bot: skip release:web,android,ios
keymanapp-test-bot
Bot
added
has-user-test
user-test-required
User Test ResultsTest specification and instructions
Results TemplateTest Artifacts |
github-actions
Bot
added
web/
web/predictive-text/
change
Base automatically changed from
change/web/remove-tokenization-corrector
to
master
September 18, 2026 13:08
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Fixes: #16577
From 14.0 up until now, our correction-search focused on the minimum "correction distance" to a possible text correction, then performed prediction on top of that. This PR changes the search to integrate word-frequency weighting from models within the search, allowing us to optimize for the most likely predictions, rather than stopping at an intermediate point. There is no additional performace cost impact of note here - the probability associated with a correction's best prediction is already accessible O(1) with operations we've already been doing.
It is likely that this will help further mitigate issues with transposition handling noted in the user tests for #16450.
Build-bot: skip release:web,android,ios
User Testing
TEST_SMOKE: Use predictive text with Keyman for Android and check for any unusual and undesired behaviors.
TEST_TEH: Using Keyman for Android with SIL EuroLatin for English, type
tehand verify that the shows up as a suggestion and is autoselected.