Add AI-powered bill classification, expense prediction, and financial analysis

[Copilot]

Implements rule-based AI system for automatic bill categorization, monthly expense forecasting, financial health analysis, and spending pattern detection.

Core Features

Bill Classification (POST /api/v1/ai/classify)

• Auto-categorizes bills via keyword matching (Chinese/English) + amount heuristics
• 10+ expense categories, 4+ income categories
• Returns confidence-scored suggestions

Expense Prediction (GET /api/v1/ai/predict)

• 6-month rolling window statistical forecasting
• Per-category breakdown with trend analysis
• 24h result caching

Financial Analysis (POST /api/v1/ai/analyze)

• Savings rate, spending trends, top categories
• Personalized insights based on thresholds (e.g., overspending warnings)

Pattern Detection (GET /api/v1/ai/patterns)

• Identifies regular vs irregular expenses by frequency analysis

Implementation

domain/ai_analysis.go              # Interfaces & models
repository/ai_analysis_repository.go  # MongoDB aggregations
usecase/ai_analysis_usecase.go     # Rule engine + statistical algorithms
controller/ai_analysis_controller.go  # HTTP handlers

Key Algorithm: Classification uses weighted keyword matching + amount ranges. Prediction applies trend multiplier to historical averages. All processing local, no external dependencies.

Example Usage:

# Classify bill
curl -X POST /api/v1/ai/classify -d '{"description":"肯德基午餐","amount":45.5}'
# Returns: {"suggestions":[{"category":"餐饮","confidence":0.85}],"type":0}

# Get predictions
curl "/api/v1/ai/predict?book_id=XXX&period=2024-02"

Testing

• 12 unit tests covering classification accuracy, trend detection, edge cases (zero division, empty data)
• CodeQL: 0 vulnerabilities

Documentation

• AI_IMPLEMENTATION_SUMMARY.md - Technical overview
• RECOMMENDATIONS.md - Usage guide addressing original requirement
• server/docs/AI_ANALYSIS_API.md - API reference
• server/docs/AI_FEATURES_CN.md - 中文说明
• server/docs/ANDROID_INTEGRATION.md - Client integration examples

Design Note: Rule-based approach provides immediate value without training data. Architecture supports future ML model drop-in replacement via interface abstraction.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• 192.168.8.6
  • Triggering command: /tmp/go-build3311262496/b472/pkg.test /tmp/go-build3311262496/b472/pkg.test -test.testlogfile=/tmp/go-build3311262496/b472/testlog.txt -test.paniconexit0 -test.timeout=10m0s -dumpdir /tmp/go-build3311262496/b441/ -dumpbase _x001.c -dumpbase-ext .c -m64 -mtune=generic (packet block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

Original prompt

基于当前项目，添加AI 分析分类账单和预测。
基于支付宝，微信，银行账单和现金归账。规划个人财务。你有什么建议

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.