Unify transform persistence for filter, sort, advQuery, and pivot

[RachanaB5]

Closes #238

Summary

This PR standardizes the behavior of view-level transformations (filter, sort, advQueryFilter, pivotTables) by making them persisted operations.

Previously, these operations could return updated data without saving to the working CSV, leading to inconsistencies between:

• Main table (UI)
• Working CSV (disk)
• Data after reload (GET /projects/get/{project_id})

Changes

Backend

• Set should_save = True for:
  • filter
  • sort
  • advQueryFilter
  • pivotTables
• Ensures transformations:
  • are written to working CSV
  • are logged consistently
• Added replay support for these operations in apply_logged_transformation

Frontend

• Filter, Sort, Advanced Query, Pivot now:
  • call onTransform after API success
  • update main table state
  • close panel after apply
• Synced ProjectContext via updateData (important for pivot/dtypes)

Tests

• Added replay unit tests (TestApplyLoggedViewTransforms)
• No integration tests added to avoid local DB/env dependency issues

Result

• Main table updates immediately after Apply
• Changes persist to working CSV
• Reload (GET) returns consistent data
• Transformations are replayable via logs/checkpoints

Notes

• Checkpoint/revert edge cases are not fully addressed here (Repeated saves drop previously checkpointed transforms by rebuilding from only pending logs #224, [Bug] Revert to original does not clear pending unapplied logs, subsequent save re-applies stale transformations #166, [Bug] Revert doesn't clear unapplied logs — saving after revert replays reverted transformations #49)
• Existing logging uses transformation_input.dict() (can migrate to model_dump() later)

How to test

1. Apply filter/sort/adv query/pivot
2. Verify table updates
3. Refresh page → data remains consistent
4. (Optional) test checkpoint/replay

Happy to split this into smaller PRs if preferred.