Lecture 26: Project Practice - Review and Summary

Review the development process of two complete projects, summarize experiences and lessons, extract reusable development patterns.

1. Project Review

1.1 Financial Intelligent Assistant

Core functions:

  • ✅ Invoice recognition and entry
  • ✅ Financial report auto-generation
  • ✅ Data processing and export
  • ✅ Intelligent Q&A support

Technical highlights:

  • OCR + regex extraction for invoice information recognition
  • Excel template + data filling for report generation
  • Conversational interaction lowers usage threshold

Challenges encountered:

  • Diverse invoice formats, recognition accuracy needs continuous optimization
  • Report templates need customization for different enterprises
  • Data security needs special attention

1.2 HR Intelligent Assistant

Core functions:

  • ✅ Resume parsing and talent pool management
  • ✅ Attendance data statistics and anomaly detection
  • ✅ Salary calculation and payslip generation
  • ✅ Employee information intelligent Q&A

Technical highlights:

  • NLP + entity extraction for resume parsing
  • Rule engine handles attendance logic
  • Formula engine supports complex salary structures

Challenges encountered:

  • Large differences in resume formats, parsing is difficult
  • Attendance rules vary by company, need flexible configuration
  • Salary calculation involves tax policies, need timely updates

2. Development Pattern Summary

2.1 General Architecture Pattern

1
2
3
4
5
Unified architecture followed by all projects:

User input → Intent recognition → Parameter extraction → Business processing → Result output
   ↓              ↓                  ↓                  ↓                ↓
Dialog interface   LLM/rules    Entity recognition    Core logic    Formatted display

2.2 Reusable Components

1. OCR Service

1
2
3
class OCRService:
    def recognize(self, image_path: str) -> str
    def extract_structured(self, image_path: str, template: dict) -> dict

2. Excel Processing

1
2
3
class ExcelProcessor:
    def fill_template(self, template_path: str, data: dict) -> str
    def read_data(self, file_path: str) -> pd.DataFrame

3. Data Storage



## 🎓 AI 编程实战课程

想系统学习 AI 编程？程序员晚枫的 **AI 编程实战课** 帮你从零上手！

- 👉 **课程报名**：[点击这里报名，前3讲免费试听](https://r7up9.xetslk.com/s/1uP5YW)
- 👉 **免费试看**：[B站免费试看前3讲，先看看适不适合自己](https://www.bilibili.com/cheese/play/ss982042944)