flyingmouse-format

`flyingmouse-format` (飞鼠格式) by LaoFeng is an offline, multi-format conversion tool built for Windows environments. Rather than leaning on cloud conversion APIs, the application operates strictly locally by bundling and orchestrating a suite of open-source CLI tools and engines: FFmpeg for media handling, LibreOffice for office documents, Poppler and Tesseract for PDF processing and OCR, `pdf2docx` for structural reconstruction, `dcraw` for camera RAW inputs, and an inline AVS3 decoder. ### Architecture & Pipeline Breakdown Under the hood, `flyingmouse-format` acts as an execution host and interface wrapper over these embedded executables: - **PDF & Document Engine**: PDF-to-Excel conversion relies on electronic text coordinates coupled with a Tesseract OCR fallback for scanned pages. It processes bordered or borderless tables, multi-page continuations, merged cells, and emits low-confidence annotations. For PDF-to-Word tasks, the embedded `pdf2docx` engine reconstructs paragraphs, fonts, and layouts on Windows 10/11 (with a plain-text extraction fallback on Windows 7). Standard PDF operations include page-level splitting, AES-256 encryption, and decryption. - **Media Transcoding**: Video exports targeting MP4, MOV, or MKV expose explicit video codec selection for H.264, H.265, and AV1. The audio pipeline handles standard unencrypted containers (MP3, WAV, FLAC, M4A, AAC, OGG, OPUS, WMA). It explicitly excludes platform-specific DRM files (such as NCM, KGG, mflac, kgma, or kwm). - **Format Conversions**: HTML and Office transformations to Markdown extract structured headers, lists, and code blocks. Image conversion supports HEIC/HEIF and TGA via FFmpeg, camera RAW files (CR2, CR3, NEF, ARW, DNG) via `dcraw`, and multi-size ICO generation. Local e-book workflows handle conversion across EPUB, MOBI, TXT, MD, and HTML formats. ### Setup and CLI Interface The project includes an Electron desktop GUI—with reactive animations tied to conversion state machine events—alongside a headless CLI (`cli.js`) and Agent skill integration for Codex and Claude catalog insertion. When building from source, note that heavy binaries within the `bin/` directory (FFmpeg, LibreOffice, Poppler, Tesseract) are omitted from the source repository and must be supplied manually prior to execution. To launch the desktop application from source: ```powershell npm install npm run desktop ``` You can execute conversion queries or batch tasks via the CLI entrypoint: ```powershell node cli.js capabilities --json node cli.js targets example.p ``` ### System Limits & Resource Protection To maintain host stability during local processing, the runtime imposes hard boundaries: - **Single Image Limit**: Max 50MP or 16384px resolution. - **Image-to-PDF Queue**: Total limit capped at 100MP (queue order can be reordered prior to generation). - **Batch Processing**: Total batch size cap set at 2GB. - **Uncapped Operations**: PDF page counts and standalone OCR tasks run without artificial page restrictions. Distributed under a strict non-commercial license, `flyingmouse-format` serves as a practical, self-contained desktop utility for complete offline file transformations. --- ### Recommended Engineering Tools If you're building modern applications, we highly recommend checking out **[Hostinger](https://hostinger.com/?aff=placeholder)** and **[Firebase](https://firebase.google.com/?aff=placeholder)** to accelerate your development workflow.