# CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

## Project Overview

This is a C# WPF application that extracts data from DWG (AutoCAD) files and processes them using AI analysis. The application has two main components:

1. **C# WPF Application** (`DwgExtractorManual`) - Main GUI application for DWG processing
2. **Python Analysis Module** (`fletimageanalysis`) - AI-powered document analysis using Gemini API

## Build and Development Commands

### C# Application
```bash
# Build the application
dotnet build

# Run the application  
dotnet run

# Clean build artifacts
dotnet clean

# Publish for deployment
dotnet publish -c Release
```

### Python Module Setup
```bash
# Run the cleanup and setup script (Windows)
cleanup_and_setup.bat

# Or manually setup Python environment
cd fletimageanalysis
python -m venv venv
call venv\Scripts\activate.bat
pip install -r requirements.txt
```

### Python CLI Usage
```bash
# Batch process files via CLI
cd fletimageanalysis
python batch_cli.py --files "file1.pdf,file2.dxf" --schema "한국도로공사" --concurrent 3 --output "results.csv"
```

## Architecture

### C# Component Structure
- **MainWindow.xaml.cs** - Main WPF window and UI logic
- **Models/DwgDataExtractor.cs** - Core DWG file processing using Teigha SDK
- **Models/ExcelDataWriter.cs** - Excel output generation using Office Interop
- **Models/TeighaServicesManager.cs** - Singleton manager for Teigha SDK lifecycle
- **Models/FieldMapper.cs** - Maps extracted data to target formats
- **Models/SettingsManager.cs** - Application configuration management

### Python Component Structure  
- **batch_cli.py** - Command-line interface for batch processing
- **multi_file_processor.py** - Orchestrates multi-file processing workflows
- **gemini_analyzer.py** - AI analysis using Google Gemini API
- **pdf_processor.py** - PDF document processing
- **dxf_processor.py** - DXF file processing
- **csv_exporter.py** - CSV output generation

### Key Dependencies
- **Teigha SDK** - DWG file reading and CAD entity processing (requires DLL files in specific path)
- **Microsoft Office Interop** - Excel file generation
- **Npgsql** - PostgreSQL database connectivity
- **Google Gemini API** - AI-powered document analysis
- **PyMuPDF** - PDF processing in Python component

## Current Development Focus

The project is undergoing a **Note Detection Refactor** (see `NoteDetectionRefactor.md`):
- Replacing fragile "horizontal search line" algorithm in `DwgDataExtractor.cs`
- Implementing robust "vertical ray-casting" approach for NOTE content box detection
- Key methods being refactored: `FindNoteBox`, `GetAllLineSegments`, `TraceBoxFromTopLine`

## Important Notes

- Teigha DLLs must be present in the specified path for DWG processing to work
- The Python module requires Google Gemini API key configuration
- Excel output uses COM Interop and requires Microsoft Office installation
- The application supports both manual GUI operation and automated batch processing via CLI