Implement server-side search endpoint for scalable file searching

[bryanchriswhite]

Summary

Currently, file search is implemented entirely client-side using Fuse.js in the React UI. While this works well for small-to-medium datasets, it has scalability limitations as the dataset grows larger.

This issue proposes implementing a server-side search endpoint to handle search queries on the backend, which will:

• Reduce client-side data transfer for large datasets
• Enable pagination of search results
• Support more advanced search features
• Improve performance for users with slower devices

Current Implementation

Client-side search in pinshare-ui/src/pages/Browse.jsx:

• Uses Fuse.js for fuzzy matching
• Searches across: fileName, ipfsCID, fileType, fileSHA256
• Field weights: fileName (0.4), ipfsCID (0.3), fileType (0.2), fileSHA256 (0.1)
• Threshold: 0.3 (fuzzy matching tolerance)
• All files are fetched via GET /files then filtered in browser

Proposed Changes

Backend API

Add new endpoint: GET /files/search

Query Parameters:

• q (required): Search query string
• limit (optional): Max results to return (default: 50)
• offset (optional): Pagination offset (default: 0)
• fields (optional): Comma-separated fields to search (default: all)
• threshold (optional): Fuzzy match threshold (default: 0.3)

Response:

{
  "results": [...],
  "total": 123,
  "limit": 50,
  "offset": 0,
  "query": "example"
}

Implementation Options

1. Go native implementation with string matching libraries
2. SQLite FTS (Full-Text Search) if/when migrating to SQLite
3. PostgreSQL full-text search if/when migrating to PostgreSQL
4. Dedicated search engine (Elasticsearch, Meilisearch, etc.) for large-scale deployments

Migration Strategy

The client-side search should remain as a fallback:

1. Try server-side search first
2. If endpoint not available (older backend), fall back to client-side Fuse.js
3. This ensures backward compatibility during rollout

Benefits

• Scalability: Handles thousands of files without performance degradation
• Reduced bandwidth: Only matching results sent to client
• Pagination: Large result sets can be paginated
• Advanced features: Can add filters, facets, aggregations, etc.
• Consistent results: Same search algorithm for all clients

Related Issues

• #TBD: PostgreSQL migration (will enable PostgreSQL full-text search)

Implementation Priority

Medium Priority - Current client-side search works well for typical use cases. This becomes important as the dataset grows beyond ~1000 files.

---

Migrated from bryanchriswhite/PinShare#2