ppu-paddle-ocr

Use this skill whenever the user is writing TypeScript/JavaScript code that imports `ppu-paddle-ocr`, `ppu-paddle-ocr/web`, or `ppu-paddle-ocr/mobile`, or asks about PaddleOCR / PP-OCRv6 / PP-OCRv5 in JavaScript runtimes - text detection, text recognition, OCR on receipts/invoices/documents, ONNX Runtime OCR, multilingual OCR (Latin, Cyrillic, Arabic, Indic, CJK, Thai), WebGPU OCR in browsers, browser-extension OCR, React Native / Expo OCR on iOS and Android (Skia + onnxruntime-react-native), INT8-quantized recognition, custom dictionaries, recognition strategies (`per-box` / `per-line` / `cross-line`), or processing engines (`opencv` / `canvas-native`). Trigger even when the user says "PaddleOCR in Node", "OCR library that runs in Bun and the browser", or "OCR in React Native" without naming the package, as long as the codebase imports it. This skill encodes the three entry points (`ppu-paddle-ocr`, `ppu-paddle-ocr/web`, and `ppu-paddle-ocr/mobile`), the lifecycle (`new` -> `initialize` -> `recognize` -> `de