Add performance optimizations and database indexes

Database Indexes (migrations 006-009):
- Add case-insensitive indexes for auth lookups (email, username)
- Add composite indexes for task kanban queries
- Add indexes for notification, document, and completion queries
- Add unique index for active share codes
- Remove redundant idx_share_code_active and idx_notification_user_sent

Repository Optimizations:
- Add FindResidenceIDsByUser() lightweight method (IDs only, no preloads)
- Optimize GetResidenceUsers() with single UNION query (was 2 queries)
- Optimize kanban completion preloads to minimal columns (id, task_id, completed_at)

Service Optimizations:
- Remove Category/Priority/Frequency preloads from task queries
- Remove summary calculations from CRUD responses (client calculates)
- Use lightweight FindResidenceIDsByUser() instead of full FindByUser()

These changes reduce database load and response times for common operations.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

migrations/009_remove_redundant_indexes.up.sql

@@ -0,0 +1,15 @@
+-- Migration: 009_remove_redundant_indexes
+-- Description: Remove indexes that are redundant or unused
+
+-- Remove redundant share code index
+-- idx_share_code_active is superseded by unique idx_sharecode_code_active (migration 008)
+-- Both filter WHERE is_active = true, but 008's unique index on (code) is more restrictive
+DROP INDEX IF EXISTS idx_share_code_active;
+
+-- Remove unused composite notification index
+-- idx_notification_user_sent on (user_id, sent) is never used:
+-- - GetPendingNotifications() only filters on sent, not user_id
+-- - FindByUser() only filters on user_id, not sent (uses idx_notification_user_created_at)
+-- - No queries combine user_id AND sent together
+-- The leading column (user_id) queries already use idx_notification_user_created_at
+DROP INDEX IF EXISTS idx_notification_user_sent;