first commit

backend/app/transcriber/base.py

@@ -0,0 +1,14 @@
+from abc import ABC, abstractmethod
+
+from app.models.transcriber_model import TranscriptResult
+
+
+class Transcriber(ABC):
+    @abstractmethod
+    def transcript(self,file_path:str)->TranscriptResult:
+        '''
+
+        :param file_path:音频路径
+        :return: 返回一个 TranscriptResult 类
+        '''
+        pass

backend/app/transcriber/transcriber_provider.py

@@ -0,0 +1,11 @@
+from app.transcriber.whisper import WhisperTranscriber
+print('实例化transcriber')
+# TODO:后面需要加入逻辑选择
+_transcriber = None
+
+def get_transcriber(model_size="base", device="cuda"):
+    global _transcriber
+    if _transcriber is None:
+        print('加载_transcriber')
+        _transcriber = WhisperTranscriber(model_size=model_size, device=device)
+    return _transcriber

backend/app/transcriber/whisper.py

@@ -0,0 +1,92 @@
+from faster_whisper import WhisperModel
+
+from app.decorators.timeit import timeit
+from app.models.transcriber_model import TranscriptSegment, TranscriptResult
+from app.transcriber.base import Transcriber
+from app.utils.env_checker import is_cuda_available, is_torch_installed
+from app.utils.path_helper import get_model_dir
+
+'''
+ Size of the model to use (tiny, tiny.en, base, base.en, small, small.en, distil-small.en, medium, medium.en, distil-medium.en, large-v1, large-v2, large-v3, large, distil-large-v2, distil-large-v3, large-v3-turbo, or turbo
+'''
+
+
+class WhisperTranscriber(Transcriber):
+    # TODO:修改为可配置
+    def __init__(
+            self,
+            model_size: str = "base",
+            device: str = 'cpu',
+            compute_type: str = None,
+            cpu_threads: int = 1,
+    ):
+        if device == 'cpu' or device is None:
+            self.device = 'cpu'
+        else:
+            self.device = "cuda" if self.is_cuda() else "cpu"
+            if device == 'cuda' and self.device == 'cpu':
+                print('没有 cuda 使用 cpu进行计算')
+
+        self.compute_type = compute_type or ("float16" if self.device == "cuda" else "int8")
+
+        model_path = get_model_dir("whisper")
+        self.model = WhisperModel(
+            model_size,
+            device=self.device,
+            # compute_type="int8",  # 或 "float16"
+            cpu_threads=cpu_threads,
+            download_root=model_path
+        )
+
+    @staticmethod
+    def is_torch_installed() -> bool:
+        try:
+            import torch
+            return True
+        except ImportError:
+            return False
+
+    @staticmethod
+    def is_cuda() -> bool:
+        try:
+            if is_cuda_available():
+                print("✅ CUDA 可用，使用 GPU")
+                return True
+            elif is_torch_installed():
+                print("⚠️ 只装了 torch，但没有 CUDA，用 CPU")
+                return False
+            else:
+                print("❌ 还没有安装 torch，请先安装")
+                return False
+
+        except ImportError:
+            return False
+
+    @timeit
+    def transcript(self, file_path: str) -> TranscriptResult:
+
+        segments_raw, info = self.model.transcribe(file_path)
+
+        segments = []
+        full_text = ""
+
+        for seg in segments_raw:
+            text = seg.text.strip()
+            full_text += text + " "
+            segments.append(TranscriptSegment(
+                start=seg.start,
+                end=seg.end,
+                text=text
+            ))
+
+        return TranscriptResult(
+            language=info.language,
+            full_text=full_text.strip(),
+            segments=segments,
+            raw=info
+        )
+
+
+if __name__ == '__main__':
+    print(WhisperTranscriber(cpu_threads=8).transcript(
+        '''D:\\data_backup_from_ssd\\02_个人项目\\11_BiliNote\\backend\\data\\BV1vcZ5YQE9X.mp3'''))