fb1e5dceba
- Upgrade PyTorch and torchaudio to 2.6.0 with CUDA 12.4 support - Update GPU reset script to gracefully stop/start Ollama via supervisorctl - Add Docker Compose configuration for both API and MCP server modes - Implement comprehensive Docker entrypoint for multi-mode deployment - Add GPU health check cleanup to prevent memory leaks - Fix transcription memory management with proper resource cleanup - Add filename security validation to prevent path traversal attacks - Include .dockerignore for optimized Docker builds - Remove deprecated supervisor configuration 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
104 lines
3.1 KiB
Docker
104 lines
3.1 KiB
Docker
# Multi-purpose Whisper Transcriptor Docker Image
|
|
# Supports both MCP Server and REST API Server modes
|
|
# Use SERVER_MODE environment variable to select: "mcp" or "api"
|
|
|
|
FROM nvidia/cuda:12.4.1-cudnn-runtime-ubuntu22.04
|
|
|
|
# Prevent interactive prompts during installation
|
|
ENV DEBIAN_FRONTEND=noninteractive
|
|
|
|
# Install system dependencies
|
|
RUN apt-get update && apt-get install -y \
|
|
software-properties-common \
|
|
curl \
|
|
&& add-apt-repository ppa:deadsnakes/ppa \
|
|
&& apt-get update && apt-get install -y \
|
|
python3.12 \
|
|
python3.12-venv \
|
|
python3.12-dev \
|
|
ffmpeg \
|
|
git \
|
|
nginx \
|
|
supervisor \
|
|
&& rm -rf /var/lib/apt/lists/*
|
|
|
|
# Make python3.12 the default
|
|
RUN update-alternatives --install /usr/bin/python python /usr/bin/python3.12 1 && \
|
|
update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.12 1
|
|
|
|
# Install pip using ensurepip (Python 3.12+ doesn't have distutils)
|
|
RUN python -m ensurepip --upgrade && \
|
|
python -m pip install --upgrade pip
|
|
|
|
# Set working directory
|
|
WORKDIR /app
|
|
|
|
# Copy requirements first for better caching
|
|
COPY requirements.txt .
|
|
|
|
# Install Python dependencies with CUDA 12.4 support
|
|
RUN pip install --no-cache-dir \
|
|
torch==2.6.0 --index-url https://download.pytorch.org/whl/cu124 && \
|
|
pip install --no-cache-dir \
|
|
torchaudio==2.6.0 --index-url https://download.pytorch.org/whl/cu124 && \
|
|
pip install --no-cache-dir \
|
|
faster-whisper \
|
|
fastapi>=0.115.0 \
|
|
uvicorn[standard]>=0.32.0 \
|
|
python-multipart>=0.0.9 \
|
|
aiofiles>=23.0.0 \
|
|
mcp[cli]>=1.2.0 \
|
|
gTTS>=2.3.0 \
|
|
pyttsx3>=2.90 \
|
|
scipy>=1.10.0 \
|
|
numpy>=1.24.0
|
|
|
|
# Copy application code
|
|
COPY src/ ./src/
|
|
COPY pyproject.toml .
|
|
|
|
# Copy test audio file for GPU health checks
|
|
COPY test.mp3 .
|
|
|
|
# Copy nginx configuration
|
|
COPY nginx/transcriptor.conf /etc/nginx/sites-available/transcriptor.conf
|
|
|
|
# Copy entrypoint script and GPU reset script
|
|
COPY docker-entrypoint.sh /docker-entrypoint.sh
|
|
COPY reset_gpu.sh /app/reset_gpu.sh
|
|
RUN chmod +x /docker-entrypoint.sh /app/reset_gpu.sh
|
|
|
|
# Create directories for models, outputs, and logs
|
|
RUN mkdir -p /models /outputs /logs /app/outputs/uploads /app/outputs/batch /app/outputs/jobs
|
|
|
|
# Set Python path
|
|
ENV PYTHONPATH=/app/src
|
|
|
|
# Default environment variables (can be overridden)
|
|
ENV WHISPER_MODEL_DIR=/models \
|
|
TRANSCRIPTION_OUTPUT_DIR=/outputs \
|
|
TRANSCRIPTION_BATCH_OUTPUT_DIR=/outputs/batch \
|
|
TRANSCRIPTION_MODEL=large-v3 \
|
|
TRANSCRIPTION_DEVICE=auto \
|
|
TRANSCRIPTION_COMPUTE_TYPE=auto \
|
|
TRANSCRIPTION_OUTPUT_FORMAT=txt \
|
|
TRANSCRIPTION_BEAM_SIZE=5 \
|
|
TRANSCRIPTION_TEMPERATURE=0.0 \
|
|
API_HOST=127.0.0.1 \
|
|
API_PORT=33767 \
|
|
JOB_QUEUE_MAX_SIZE=5 \
|
|
JOB_METADATA_DIR=/outputs/jobs \
|
|
JOB_RETENTION_DAYS=7 \
|
|
GPU_HEALTH_CHECK_ENABLED=true \
|
|
GPU_HEALTH_CHECK_INTERVAL_MINUTES=10 \
|
|
GPU_HEALTH_TEST_MODEL=tiny \
|
|
GPU_HEALTH_TEST_AUDIO=/test-audio/test.mp3 \
|
|
GPU_RESET_COOLDOWN_MINUTES=5 \
|
|
SERVER_MODE=api
|
|
|
|
# Expose port 80 for nginx (API mode only)
|
|
EXPOSE 80
|
|
|
|
# Use entrypoint script to handle different server modes
|
|
ENTRYPOINT ["/docker-entrypoint.sh"]
|