Reduce redundant dataset fetches and standardize notebook API endpoints

[Irozuku]

Summary

This pull request standardizes notebook-related API endpoint formatting by ensuring consistent trailing slash usage across notebook API calls. It also improves dataset refresh behavior by only triggering dataset fetches when dataset jobs are finished, reducing unnecessary API requests and redundant dataset reloads.

From the image (running without React StrictMode), only the three necessary fetches are performed with these changes:

• Datasets to display on the left sidebar
• Notebooks to display on the left sidebar
• Jobs

---

Type of Change

Check all that apply like this [x]:

• Backend change
• Frontend change
• CI / Workflow change
• Build / Packaging change
• Bug fix
• Documentation

---

Changes (by file)

Briefly list the important modified files and what was done.

• notebook.ts: Updated notebookEndpoint to consistently use a trailing slash and adjusted notebook-related API calls to avoid malformed URLs or double slashes.
• useDatasets.ts: Updated job subscription logic to only trigger fetchDatasets when a DatasetJob finishes.
• DatasetsAndNotebooksProvider.tsx: Removed the initial dataset fetch on mount so datasets are refreshed only when relevant finished jobs are received.

---

Testing (optional)

• Verify notebook create, retrieve, update, and delete operations work correctly with the updated endpoint formatting.
• Confirm explorer and converter requests by notebook ID still resolve correctly.
• Trigger dataset jobs and verify datasets are refreshed only after jobs finish.
• Confirm unnecessary dataset fetches are no longer triggered during intermediate job states.

---

Notes (optional)

• These changes reduce redundant frontend API calls.