()
| 30 | |
| 31 | |
| 32 | def get_engine(): |
| 33 | global _engine |
| 34 | if _engine is not None: |
| 35 | return _engine |
| 36 | if RapidOCR is None: |
| 37 | raise RuntimeError('缺少 rapidocr 依赖，无法使用 catpcha_v2 识别算法。') |
| 38 | |
| 39 | _engine = RapidOCR( |
| 40 | params={ |
| 41 | 'Det.engine_type': EngineType.ONNXRUNTIME, |
| 42 | 'Det.lang_type': LangDet.CH, |
| 43 | 'Det.model_type': ModelType.MOBILE, |
| 44 | 'Det.ocr_version': OCRVersion.PPOCRV5, |
| 45 | 'Rec.engine_type': EngineType.ONNXRUNTIME, |
| 46 | 'Rec.lang_type': LangRec.CH, |
| 47 | 'Rec.model_type': ModelType.MOBILE, |
| 48 | 'Rec.ocr_version': OCRVersion.PPOCRV5, |
| 49 | } |
| 50 | ) |
| 51 | return _engine |
| 52 | |
| 53 | |
| 54 | def normalize_text(text): |
| 55 | return (text or '').replace(' ', '').replace('\n', '').strip() |
| 56 | |
| 57 | |
| 58 | def extract_chinese_chars(text): |
| 59 | return [char for char in normalize_text(text) if '\u4e00' <= char <= '\u9fff'] |
| 60 | |
| 61 |
no test coverage detected