-
Notifications
You must be signed in to change notification settings - Fork 18
Open
Milestone
Description
Determine the work required for the ml-trainer equivalent makecode integration
Investigation
- Add the makecode-embed library
- Determine if it should be embedded within the 'mainContent' window or it should be a page by itself
- (If it should take up the entire page, like with ML-Trainer) Create a exceptional route in the router that fills the entire page
- Create a MakeCodeProject from the current state of the application
- Create an "extension" that can be used inside of the embedded makecode, there's probably something we can borrow from ML-Trainer
- Use the block rendering to display the configuration on the model page instead of the current threshold logic
Implementation
See the dev-makecode-embed branch for reference implementations
- Add the makecode-embed library
- Add /makecode path
- Add ml machine extension
- Insert makecode embed iframe
- Move the makecode frame driver to controller
- WontDo - it's too tied to the UI
- Alternative, create a FrameBuilder or some other way to move the native js away from Svelte framework code
- Add back button to embedded makecode
- Move navigation into backend interface, replace in makecode controller
- Leave it in the frontend and pass in event handlers to the frambuilder
- Add 'to makecode' button on output page
- Put FeatureToggle in interface-controller
- Load the features when ML-Machine is instantiated
- Add makecode override state
- Add blocks renderer to makecode output page view
- Save MakeCode project
- Create MakeCodeRepository
- Create MakeCodeService
- Load saved MakeCode project from repository
Potential further work (Solution for #637)
- Generate project files instead of using extension
Metadata
Metadata
Assignees
Labels
No labels