dockerfile

Dockerfile

@@ -0,0 +1,20 @@
+FROM python:3.8
+
+RUN apt-get update -y
+RUN apt-get install ffmpeg -y
+RUN apt-get install vim -y
+RUN apt-get update -y
+RUN apt-get install fonts-noto-cjk fonts-noto-color-emoji
+
+RUN pip install --upgrade pip
+
+WORKDIR /app
+
+ADD ./requirements.txt /app/requirements.txt
+RUN pip install --no-cache-dir --upgrade -r requirements.txt
+
+ADD ./ /app/
+
+RUN mkdir -p /usr/local/var/log
+
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]

docker-compose.yml

@@ -0,0 +1,11 @@
+version: "3"
+services:
+  tts_sambert_hifigan:
+    build: .
+    command: python -m uvicorn main:app --host 0.0.0.0 --port 8001 --reload
+    restart: always
+    volumes:
+      - .:/app
+      - /etc/localtime:/etc/localtime # 同步时间
+    ports:
+      - "8001:8001"

main.py

@@ -0,0 +1,32 @@
+import os
+
+from typing import Union
+
+from fastapi import FastAPI
+from fastapi.responses import FileResponse
+from modelscope.outputs import OutputKeys
+from modelscope.pipelines import pipeline
+from modelscope.utils.constant import Tasks
+from datetime import datetime
+
+model_id = "damo/speech_sambert-hifigan_tts_zhitian_emo_zh-cn_16k"
+sambert_hifigan_tts = pipeline(task=Tasks.text_to_speech, model=model_id)
+
+app = FastAPI()
+
+
+def generate_audio(text: str):
+    output = sambert_hifigan_tts(input=text)
+    wav = output[OutputKeys.OUTPUT_WAV]
+    timestamp = datetime.now().strftime("%Y%m%d%H%M%S")  # Generate timestamp
+    filename = f"media/audio/{timestamp}.wav"  # Use timestamp as filename
+    os.makedirs(os.path.dirname(filename), exist_ok=True)  # Create directories if they don't exist
+    with open(filename, "wb") as f:
+        f.write(wav)
+    return filename
+
+
+@app.get("/tts", responses={200: {"content": {"audio/wav": {}}}})
+def get_tts(text: Union[str, None] = None):
+    filename = generate_audio(text)
+    return FileResponse(filename, filename=f"{os.path.basename(filename)}")

requirements.txt

@@ -0,0 +1,4 @@
+-f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html
+modelscope[audio]
+fastapi
+uvicorn[standard]