Fix: Install cmake and build-essential for llama-cpp-python build

Dockerfile

@@ -2,23 +2,21 @@ FROM python:3.12-slim
 
 WORKDIR /app
 
-# Install system dependencies
+# Install system dependencies for building llama-cpp-python
 RUN apt-get update && apt-get install -y \
-    gcc \
-    g++ \
-    make \
+    build-essential \
+    cmake \
+    git \
     && rm -rf /var/lib/apt/lists/*
 
 # Copy requirements and install
 COPY requirements.txt .
 
-# Set environment variables to use pre-built wheels for llama-cpp-python
-ENV CMAKE_ARGS="-DLLAMA_BLAS=OFF -DLLAMA_BLAS_VENDOR=OFF -DLLAMA_CUBLAS=OFF"
-ENV FORCE_CMAKE=0
+# Set environment variables for CPU-only build (no BLAS, no CUDA)
+ENV CMAKE_ARGS="-DLLAMA_BLAS=OFF -DLLAMA_CUBLAS=OFF -DLLAMA_METAL=OFF"
 
-# Install llama-cpp-python from PyPI (not building from source)
-RUN pip install --no-cache-dir --only-binary=:all: llama-cpp-python==0.3.14 || \
-    pip install --no-cache-dir llama-cpp-python==0.3.14
+# Install llama-cpp-python (will build from source with CPU-only)
+RUN pip install --no-cache-dir llama-cpp-python==0.3.14 --verbose
 
 # Install remaining requirements
 RUN pip install --no-cache-dir -r requirements.txt