updated dockerfile for quantized

Dockerfile

@@ -1,38 +1,35 @@
-# Use official Python 3.11 image
+# Use official Python 3.11 base image
 FROM python:3.11
 
-# Install system dependencies needed for building Python packages
+# Install system dependencies
 RUN apt-get update && apt-get install -y \
     build-essential \
+    cmake \
+    git \
+    curl \
     && rm -rf /var/lib/apt/lists/*
 
-# Set working directory inside the container
+# Set working directory
 WORKDIR /app
 
-# Set Hugging Face cache directory and grant permissions
-# This helps with model downloads and caching within the Space
+# Set Hugging Face and ChromaDB cache directories
 ENV TRANSFORMERS_CACHE=/app/.cache \
-    HF_HOME=/app/.cache
-RUN mkdir -p /app/.cache && chmod -R 777 /app/.cache
-
-# ✅ Ensure ChromaDB can write its persistent database
-# This directory will hold the DB built at runtime.
-# It MUST be a consistent, writable location for persistence.
-RUN mkdir -p /app/vector_database && chmod -R 777 /app/vector_database
+    HF_HOME=/app/.cache \
+    PYTHONUNBUFFERED=1 \
+    PERSIST_DIR=/app/vector_database
 
-# Copy only the requirements file to leverage Docker cache
-COPY requirements.txt .
+RUN mkdir -p /app/.cache && chmod -R 777 /app/.cache
+RUN mkdir -p ${PERSIST_DIR} && chmod -R 777 ${PERSIST_DIR}
 
 # Install Python dependencies
+COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
-# Copy the rest of your application code, including 'app/' directory and 'processed_chunks.json'
-# Assuming 'app' and 'processed_chunks.json' are at the root level of your project
+# Copy project code
 COPY . .
 
-# Expose the port the app runs on
+# Expose the port FastAPI will run on
 EXPOSE 7860
 
-# Command to run the FastAPI application
-# 'app.app' refers to the 'app' FastAPI instance within 'app.py' inside the 'app' package
-CMD ["uvicorn", "app.app:app", "--host", "0.0.0.0", "--port", "7860"]
+# Run the FastAPI app using uvicorn
+CMD ["uvicorn", "app.app:app", "--host", "0.0.0.0", "--port", "7860"]