Add proof phase frontend integration

[amoghbanta]

Summary

• Wire the frontend to the proof phase backend (merged in Add image captioning proof step with per-page LLM batching #40) for post-storyboard image captioning
• Add SSE-based real-time progress tracking with polling fallback, reusing the same event format as the pipeline
• Display generated image captions (with collapsible reasoning) in the page detail view
• Replace language chip grid with searchable dropdown, add Sinhala and Tamil (fixes Add book third step is not seen completely - 14'' Laptop with 1920x1080 - 150% scale #47)

Changes

Commit	Description
1	API client types and methods (ProofStatus, runProof, getProofStatus, imageCaptioning on PageDetail)
2	Extend StepName union with "image-captioning", add label (not added to STEP_ORDER — proof-only)
3	New use-proof.ts — useProofSSE, useRunProof, useProofStatus hooks
4	New ProofProgress card component with single step indicator
5	Wire proof hooks and UI into book detail page (shown when storyboard accepted)
6	ImageCaptionList component in page detail with thumbnails and collapsible reasoning
7	Add image-captioning to debug LLM logs tab
8	Replace language picker with searchable dropdown, add Sinhala + Tamil, fix small viewport overflow (#47)

Test plan

• pnpm test — all existing tests pass
• Open a book with accepted storyboard → proof section visible
• Click "Run Proof" → SSE progress streams, step indicator animates
• After completion → page detail shows captions under images
• Refresh mid-proof → SSE auto-reconnects
• Re-run proof after completion → progress resets and streams again
• Debug panel → LLM logs tab shows image-captioning entries
• Add book wizard → step 3 language dropdowns are searchable, Sinhala + Tamil present
• Add book wizard → buttons reachable on small viewports (14" laptop, 150% scale)