fix(ui): keep base model selection consistent across training flows

[Marcelo (madpsys)]

Description

Fixes the Base Model dropdown in the training form so that:

• The selected base model remains visible after selection.
• Existing models recover their selected catalog entry from initialWeightsUrl.
• Catalog requests omit absent filters instead of sending eventTypes=undefined.
• Training launched from the Labeling Tool and project listing uses the same event-type and imagery-source filters.
• Catalog entries with null descriptions no longer cause an error.
• External catalog entries without a modelId use baseModelName as a stable fallback key.
• Removes unused leftPanel and rightPanel components.

This change also removes the unused legacy Labeling Tool panel components.

Fixes #

Type of change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that changes existing behavior)
• Documentation update
• Infrastructure / CI change

Checklist

• I have read CONTRIBUTING.md
• My changes follow the project's coding standards (PEP 8 for Python, ESLint rules for JS/TS)
• I have added or updated tests that cover my changes
• Python tests pass locally (cd hastelib && hatch run test:pytest) and the UI lints clean (cd ui && npm run lint)
• I have updated the relevant documentation (README, docs/, inline comments)
• I have added an entry to CHANGELOG.md if this is a user-facing change

Testing

The following validations were completed:

• node --test
  • 100 tests passed.
• node --test src/Components/BaseModelDropdownHelper.test.js
  • 6 focused regression tests passed.
• npm run build
  • Production build completed successfully.
• git diff --check
  • No whitespace errors found.
• Targeted ESLint validation passed for the new and modified dropdown helper files.

The repository-wide UI lint still reports pre-existing errors unrelated to this change, so the combined Python/UI validation checklist remains unchecked.

Additional context

Training consumes initialWeightsUrl, so the submitted training payload remains compatible with the existing backend contract. baseModelId is maintained in UI state to control the dropdown and is resolved from the catalog checkpoint URL when editing an existing model.

The authentication configuration for the Model Catalog routes was not changed. Those routes continue to use AuthLevel.FUNCTION.

[Copilot]

Pull request overview

Aligns base-model selection and filtering across UI training flows.

Changes:

• Adds catalog query, normalization, and selection helpers with regression tests.
• Restores selections from checkpoint URLs and safely handles null descriptions.
• Passes project event types and image-layer metadata from the Labeling Tool.

Reviewed changes

Copilot reviewed 9 out of 9 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
BaseModelDropdown.jsx	Uses normalized, controlled selections.
BaseModelDropdownHelper.js	Adds catalog and selection utilities.
BaseModelDropdownHelper.test.js	Tests helper behavior.
CreateEditModelTrainingHelper.js	Builds filtered catalog options.
CreateEditModelTrainingModal.jsx	Restores the selected base model.
LabelingTool/LabelingTool.jsx	Loads training filter metadata.
LabelingTool/LabelingToolRightPanel.jsx	Passes consistent training filters.
LabelingTool/LeftPanel.jsx	Removes obsolete panel.
LabelingTool/RightPanel.jsx	Removes obsolete panel.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.