Dockerfile

FROM ubuntu:22.04

ENV DEBIAN_FRONTEND=noninteractive
RUN apt-get update && apt-get install -y \
    curl ca-certificates python3 python3-pip bash && \
    rm -rf /var/lib/apt/lists/*

# Install Ollama
RUN curl -fsSL https://ollama.com/install.sh | sh

WORKDIR /app
COPY requirements.txt /app/requirements.txt
RUN pip3 install --no-cache-dir -r /app/requirements.txt

COPY app.py /app/app.py
COPY start.sh /app/start.sh
RUN chmod +x /app/start.sh

# Ollama
ENV OLLAMA_HOST=0.0.0.0:11434
# CPU-friendly
ENV OLLAMA_CONTEXT_LENGTH=4096
ENV OLLAMA_NUM_PARALLEL=1
ENV OLLAMA_MAX_LOADED_MODELS=1
ENV OLLAMA_KEEP_ALIVE=10m

# For Gradio on Spaces
ENV GRADIO_SERVER_NAME=0.0.0.0
ENV GRADIO_SERVER_PORT=7860

# App defaults (можно переопределить в Space Variables)
ENV MODEL_NAME=qwen2.5-coder:3b
ENV NUM_CTX=4096
ENV MAX_TOKENS=1024
ENV LITELLM_TIMEOUT=3600

CMD ["/app/start.sh"]