I want advice on Implementing Predictive Text Features in Keyman

Hello Community Member,

I am working on a custom keyboard layout for a minority language & I m exploring the possibility of incorporating predictive text features to enhance the typing experience for users; I have managed to create a basic layout but integrating predictive text is proving to be more challenging than I anticipated.

What are the best approaches to implementing predictive text in a Keyman keyboard? Is there a specific framework or tool within Keyman that supports this feature or would I need to integrate an external solution??

Since this language does not have a widely available corpus; I am thinking about creating a custom language model for predictive text. Could anyone provide guidance on how to start this process? Any tips on collecting and preparing data for this purpose would be extremely helpful.

How does adding predictive text impact the performance of the keyboard on mobile devices?

Also i have read this resorse/artical; https://community.software.sil.org/t/need-guidance-on-customizing-keyboard-layouts-in-keyman-developeraws-devops-certibut have not found any solution still need your advice. if you have any advice It woud greatly appreciate

Thank you… :blush:

Welcome @mirellehgf

Lots of good questions there. I recommend reading through some of the guides starting with:
https://help.keyman.com/developer/17.0/guides/lexical-models/intro/

which can help answer some of your questions.

Essentially with a wordlist, Keyman Developer can create a lexical model package which can then be installed with Keyman for Android or Keyman for iPhone and iPad. When the mobile device uses a Keyman keyboard with a matching language tag as the lexical-model, you’ll get suggestions above the keyboard.

1 Like

I would also ask how isolating, inflectional/agglutinating or fusional is the language morphology and writing system?