# CLAUDE.md

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

## Solution Structure

Three projects in `ComicCompress.sln`, all targeting .NET 8.0:

- **ComicCompress** (this directory) — Console application entry point. `Program.cs` handles CLI parsing via `McMaster.Extensions.CommandLineUtils` and orchestrates processing.
- **iMGcompress-core** (`../iMGcompress-core`) — Class library containing all compression logic. Both the CLI and GUI depend on this.
- **iMGcompress** (`../iMGcompress`) — Windows Forms GUI (`net8.0-windows`) wrapping the core library.

## Build & Run

```powershell
# Development build
dotnet build

# Run CLI (from project root)
dotnet run -- --help
dotnet run -- -i <input> -o <output> -q 75 -p

# Publish self-contained Windows executable
dotnet publish -r win-x64 -c Release
```

`launchSettings.json` defines named profiles (e.g. `PROD`, `DEV_HOME`, `Compress Image to webp`) for common invocations.

## CLI Flags

| Flag | Description |
|------|-------------|
| `-i` | Input file or folder (required) |
| `-o` | Output base path (default: `converted_comics`) |
| `-r` | Recurse into subfolders |
| `-s` | Skip if output already exists |
| `-q` | WebP quality 0–100 (default: 75) |
| `-p` | Parallel processing (all CPUs) |
| `-c` | Parallel with N CPUs |
| `-d` | PDF-only mode |
| `-j` | Image-only mode |
| `-n` | Zebra PRN mode (generates PNG + WebP Base64 URI) |
| `-a` | Copy unprocessed files to output |

## Architecture

**Data flow:** Input (`.cbr`/`.cbz` archives or loose images) → extract → convert each image to WebP → repack as `.cbz`.

### Core library (`iMGcompress-core`)

- **`Compressor.cs`** — Central engine. Reads RAR/ZIP archives via `SharpCompress`, converts images to WebP via `Imazen.WebP`, and writes output `.cbz`. Handles solid and non-solid archives separately. The first image in an archive (cover) is preserved without conversion.
- **`FileProcessor.cs`** — Orchestrates parallel or sequential processing over a file list. Inner class `ProcessorFlags` carries all options. Writes the final compression report.
- **`FileParser.cs`** — Recursively enumerates input files.
- **`ImageExtractor.cs`** — Implements `IRenderListener` (iText) to extract images from PDF pages and produce a `.cbz`.
- **`PixelMap.cs`** — Parses Portable Bitmap formats (PBM/PGM/PPM, variants P1–P6) into `System.Drawing.Bitmap` before WebP encoding.
- **`Logger.cs`** — Thread-safe colored console/file logger. `LogLevel`: Error, Warning, Info, Verbose.
- **`file_stat.cs`** — Simple DTO tracking per-file uncompressed/compressed sizes for statistics.

### Console app (`ComicCompress`)

`Program.cs` parses flags, builds a `ProcessorFlags`, calls `FileParser` to enumerate targets, then hands off to `FileProcessor`. Video (MP4→WebM) conversion uses `Xabe.FFmpeg`. Zebra PRN rendering uses `BinaryKits.Zpl.Viewer`.

## Key Dependencies

| Package | Purpose |
|---------|---------|
| `SharpCompress` | Read/write RAR and ZIP archives |
| `Imazen.WebP` + `libwebp.dll` | WebP encoding (requires `libwebp.dll` in output dir) |
| `Magick.NET-Q8-x64` | Advanced image format support (TIFF, etc.) |
| `iTextSharp` / `itext` | PDF parsing |
| `Xabe.FFmpeg` | Video conversion |
| `BinaryKits.Zpl.Viewer` | Zebra PRN label rendering |
| `McMaster.Extensions.CommandLineUtils` | CLI argument parsing |

**`libwebp.dll` must be present** alongside the executable at runtime — it is checked into the repo root.