Ai studio prototype

[dju90]

NOT MEANT TO BE MERGED TO PROD - only for pulling the branch to see it for yourself

To run the frontend folder, I used:

cd code-dot-org/frontend
corepack enable
yarn install

cd packages/labs/ailab
yarn dev

This PR changes AI Lab from one linear flow to an algorithm-first workflow with tabs. Students can choose KNN or Decision Tree before they select data. They can explore data, train a model, test it, and export it in separate sections.

Changes

• Add an algorithm selection screen.
• Add ID3 Decision Tree training and prediction.
• Add support for saved ID3 models in App Lab prediction code.
• Add a top navigation rail for Dataset, Train, Test, and Export.
• Show the selected algorithm in the rail.
• Add a restart warning when the user clicks the algorithm label.
• Move label and feature selection from Dataset to Train.
• Add table and card views for dataset exploration.
• Let users click a test result row to fill the Try it out form.
• Add richer column charts, including histograms, box plots, and mixed relationship plots.
• Add an Export step with model card fields.
• Show a Blockly-style getPrediction(data) preview in Export.
• Add a Decision Tree visualization.
• Add step-by-step path trace for Decision Tree predictions.
• Show leaf accuracy from accuracy-check rows.
• Add tooltips for decision nodes with split usefulness, sample count, and feature contribution.

Notes

This is prototype work for the AI Lab student flow. The main risk is UI complexity in the Decision Tree view for large trees.