feat: Smart weekly financial digest with trends & insights (bounty #121)

[alexanderxfgl-bit]

Bounty Claim: $500 — Issue #121

Implementation

Production-ready smart digest system with weekly/monthly financial summaries, trend detection, and auto-generated insights.

Backend Changes

Model (app/models.py):

• WeeklyDigest: user_id, period_start, period_end, period_type, total_spent, category_breakdown (JSON), trends (JSON), insights (JSON), generated_at

Service (app/services/digest.py):

• generate_digest() — aggregate expenses by category for current/previous period
• _category_breakdown() — convert query results to {category: total} dict
• _detect_trends() — compare current vs previous, flag >15% changes as significant
• _generate_insights() — produce human-readable insights: top category, spending spikes, overall change
• _period_bounds() — compute week/month date ranges (Monday-based weeks, 1st-based months)
• Upsert logic — updates existing digest or creates new one
• get_digest_history() — paginated historical digest listing with period filter

Routes (app/routes/digest.py):

• GET /digest?period=weekly|monthly — generate/retrieve current period digest
• GET /digest/history?limit=10&period=weekly — list historical digests
• POST /digest/generate — force generate for specific period/date

Schema (app/db/schema.sql):

• weekly_digests table with JSONB columns for breakdown/trends/insights
• Unique index on (user_id, period_start, period_end) for upsert

Tests

11 tests covering:

• Weekly and monthly digest generation
• Invalid period parameter validation
• Digest with actual expenses (category breakdown)
• History endpoint with limit and period filters
• Force generation with reference date
• Invalid date/period validation
• Trend detection for spending changes

Closes #121