go

Dockerfile

@@ -4,37 +4,61 @@ ENV PYTHONDONTWRITEBYTECODE=1 \
     PYTHONUNBUFFERED=1 \
     PORT=7860 \
     HF_HOME=/app/cache \
+    CPU_THREADS=2 \
     PATH="/home/user/.local/bin:${PATH}"
 
 WORKDIR /app
 
-# Install build dependencies
+# Install system dependencies
+# Note: We kept build-essential/cmake just in case, but they aren't strictly needed for pre-built wheels
 RUN apt-get update && apt-get install -y \
     build-essential \
     cmake \
     curl \
     git \
+    libgomp1 \
     && rm -rf /var/lib/apt/lists/*
 
-# Create user
-RUN useradd -m -u 1000 user
-RUN mkdir -p /app/cache /app/models && chown -R user:user /app
+# Create non-root user
+RUN useradd -m -u 1000 user && \
+    mkdir -p /app/cache /app/models && \
+    chown -R user:user /app
 
-# Install pip as root
+# Upgrade pip as root
 RUN pip install --no-cache-dir --upgrade pip
 
+# Switch to non-root user
 USER user
 
-# Build and install llama-cpp-python with proper flags
-RUN CMAKE_ARGS="-DGGML_BLAS=OFF -DGGML_NATIVE=OFF" \
-    pip install --no-cache-dir llama-cpp-python==0.3.2
+# ---------------------------------------------------------------------------
+# FIX: Use pre-built wheels to avoid "Timed out wheel creation"
+# This installs the latest version compatible with SmolVLM without compiling
+# ---------------------------------------------------------------------------
+RUN pip install --no-cache-dir --upgrade llama-cpp-python \
+    --extra-index-url https://abetlen.github.io/llama-cpp-python/whl/cpu
 
-# Install other dependencies
+# Copy requirements and install dependencies
 COPY --chown=user:user requirements.txt .
-RUN pip install --no-cache-dir -r requirements.txt
+# Remove llama-cpp-python from requirements.txt if it's there to prevent re-installation attempts
+RUN sed -i '/llama-cpp-python/d' requirements.txt && \
+    pip install --no-cache-dir --user -r requirements.txt
 
-# Copy app
+# Copy application structure
+COPY --chown=user:user config.py .
 COPY --chown=user:user main.py .
+COPY --chown=user:user models/ ./models/
+COPY --chown=user:user services/ ./services/
+COPY --chown=user:user routers/ ./routers/
+COPY --chown=user:user utils/ ./utils/
+
+# Create __init__.py files if they don't exist
+RUN touch models/__init__.py services/__init__.py routers/__init__.py utils/__init__.py
+
+# Health check
+HEALTHCHECK --interval=30s --timeout=10s --start-period=60s --retries=3 \
+    CMD curl -f http://localhost:7860/ping || exit 1
 
 EXPOSE 7860
-CMD ["bash", "-c", "while true; do curl -s https://xce009-ai-chat-api.hf.space/ping > /dev/null || true; sleep 300; done & python -m uvicorn main:app --host 0.0.0.0 --port 7860"]
+
+# Production startup
+CMD ["bash", "-c", "while true; do curl -s http://localhost:7860/ping > /dev/null 2>&1 || true; sleep 300; done & exec python -m uvicorn main:app --host 0.0.0.0 --port 7860 --log-level info"]