frontend: show backend validation details for 400 errors

[Bhawna109]

This PR fixes issue #400 by addressing both parts of the problem:

1. Frontend showed a generic 400 toast without backend validation details.
2. Backend had conflicting ranking constraints that could make ranking submission fail with 400.

Frontend Fix

Updated frontend/src/services/alertService.js to prioritize backend error payloads:

1. response.data.errors (array or string)
2. response.data.message + response.data.detail
3. fallback to Axios error.message

Result: users now see actionable backend validation messages instead of only:
Request failed with status code 400.

Backend Fix

Updated montage/juror_endpoints.py:

1. In get_tasks_from_round(), ranking rounds now fetch all open tasks using total_open_tasks (not fixed 100).
2. In submit_ratings(), the MAX_RATINGS_SUBMIT limit is now enforced only for rating and yesno, not ranking.

Result: ranking rounds with more than 100 entries can submit full ballots as required by backend ranking validation.

Scope

• Frontend + backend change.
• No schema changes.
• No behavior change for rating / yesno limits.

Testing

1. Triggered 400 flow and verified frontend now shows backend validation text.
2. Verified backend module compiles after changes:
   • python -m py_compile montage/juror_endpoints.py

Production Verification Note

I attempted to verify on the linked Toolforge campaign from issue #400, but my account was redirected to #/permission-denied for that route.
So I could not fully execute the affected ranking flow on production data.

Please verify this PR on the original campaign/round with coordinator/juror permissions.