Dockerfile.ufpalign

FROM kaldiasr/kaldi:latest

LABEL maintainer="UFPAlign Docker Implementation"
LABEL description="UFPAlign - Brazilian Portuguese Forced Phonetic Alignment Tool"
LABEL version="1.0"

# Set environment variables
ENV UFPALIGN_DIR=/opt/UFPAlign
ENV KALDI_ROOT=/opt/kaldi
ENV LC_ALL=pt_BR.UTF-8
ENV LANG=pt_BR.UTF-8
ENV PYTHONPATH=/opt/UFPAlign:$PYTHONPATH

# Update system and install dependencies
RUN apt-get update && \
    apt-get install -y --no-install-recommends \
    sudo \
    curl \
    wget \
    openjdk-8-jdk \
    locales \
    python3-pip \
    python3-dev \
    python3-setuptools \
    build-essential \
    sox \
    ffmpeg \
    git && \
    # Configure locale for Portuguese (Brazil)
    sed -i '/pt_BR.UTF-8/s/^# //g' /etc/locale.gen && \
    locale-gen && \
    # Upgrade pip
    pip3 install --upgrade pip && \
    # Create UFPAlign directory
    mkdir -p $UFPALIGN_DIR && \
    # Clean up apt cache
    apt-get clean && \
    rm -rf /var/lib/apt/lists/*

# Install Python dependencies
RUN pip3 install --no-cache-dir \
    gdown==5.0.0 \
    numpy==1.26.3 \
    pandas==2.2.2 \
    TextGrid==1.5 \
    Unidecode==1.3.8 \
    fastapi \
    uvicorn \
    python-multipart \
    pydantic \
    aiofiles

# Clone UFPAlign repository
RUN cd /opt && \
    git clone https://github.com/falabrasil/ufpalign.git UFPAlign && \
    cd UFPAlign && \
    chmod +x ufpalign.sh

# Set working directory
WORKDIR $UFPALIGN_DIR

# Copy application files if they exist in build context
COPY app.py /opt/UFPAlign/ 2>/dev/null || true
COPY requirements.txt /opt/UFPAlign/ 2>/dev/null || true

# Create necessary directories
RUN mkdir -p \
    $UFPALIGN_DIR/uploads \
    $UFPALIGN_DIR/output \
    $UFPALIGN_DIR/textgrid \
    /root/logs

# Download models during build (optional - can be done at runtime)
# RUN utils/download_model.sh "data" $UFPALIGN_DIR && \
#     utils/download_model.sh "mono" $UFPALIGN_DIR

# Make sure Kaldi is properly linked
RUN if [ -d "/opt/kaldi" ]; then \
        ln -sf /opt/kaldi/egs/wsj/s5/steps $UFPALIGN_DIR/steps; \
        ln -sf /opt/kaldi/egs/wsj/s5/utils $UFPALIGN_DIR/utils_kaldi; \
        ln -sf /opt/kaldi/egs/wsj/s5/path.sh $UFPALIGN_DIR/path.sh; \
    fi

# Set proper permissions
RUN chown -R root:root $UFPALIGN_DIR && \
    chmod -R 755 $UFPALIGN_DIR

# Expose port for API if needed
EXPOSE 7860

# Health check
HEALTHCHECK --interval=30s --timeout=30s --start-period=60s --retries=3 \
    CMD [ -f "$UFPALIGN_DIR/ufpalign.sh" ] && echo "UFPAlign is ready" || exit 1

# Default command
CMD ["/bin/bash"]