Remove the legacy editor and run view

[Mbeaulne]

Description

Removes the legacy editor and run view now that the current experiences are the defaults.

This change:

• Removes the v2_editor beta flag and editor/run version toggles
• Routes newly opened pipelines and runs to the current editor and run view
• Redirects legacy editor and run URLs while preserving pipeline names, run IDs, subgraph execution IDs, and run search parameters
• Deletes the legacy editor, run view, canvas, providers, hooks, and utilities that are no longer used
• Removes the unused @dnd-kit/core dependency
• Updates shared types, route helpers, unit tests, and Playwright helpers for the current implementations

Related Issue and Pull requests

• Enable the published component library for all users  #2669

Type of Change

• Improvement
• Cleanup/Refactor

Checklist

• I have tested this does not break current pipelines / runs functionality
• I have tested the changes on staging

Screenshots (if applicable)

Not applicable.

Test Instructions

1. Create a pipeline and confirm it opens in the current editor.
2. Open an existing pipeline and verify editing, saving, task connections, and task deletion.
3. Open a run and a nested subgraph execution and confirm each loads in the current run view.
4. Visit legacy /editor/:name, /runs/:id, and /runs/:id/:subgraphExecutionId URLs and confirm they redirect to their current equivalents.
5. Verify run URL search parameters are preserved after redirection.
6. Confirm the editor and run version toggles no longer appear in the application menu or beta settings.

[github-actions]

🎩 Preview

A preview build has been created at: remove_v2_editor_flag/02154c3

[camielvs]

I think we need to be careful about how we approach this task. Even this week I am still finding things that are missing from V2, and we've heard recently that people are still using V1.

I am in favour of a staged removal. We should start by making V2 the default loaded onto the base editor and runs routes, monitor for a week or two to ensure user-adoption, stability and feature-completeness, and then finally remove V1 altogether.

Staging it in pieces also affords us better reversibility if something goes wrong.