Update Dockerfile

Dockerfile

@@ -1,17 +1,16 @@
 # Base image with Python and common dependencies
 FROM python:3.10-slim
 
-# Set working directory
 WORKDIR /app
 
-# Set environment variables
 ENV DEBIAN_FRONTEND=noninteractive \
     PYTHONUNBUFFERED=1 \
     PYTHONDONTWRITEBYTECODE=1 \
     TRANSFORMERS_CACHE=/app/cache \
-    HF_HOME=/app/cache
+    HF_HOME=/app/cache \
+    NLTK_DATA=/app/nltk_data \
+    MPLCONFIGDIR=/app/.config/matplotlib
 
-# COPY requirements file and install Python dependencies
 COPY requirements.txt requirements.txt 
 COPY extract_img_pdf.py extract_img_pdf.py
 COPY live_streaming_flask.py live_streaming_flask.py
@@ -19,6 +18,7 @@ COPY templates/ /app/templates
 COPY .env .env
 COPY test_streaming.py test_streaming.py 
 COPY app_main.py app_main.py
+COPY compare_image_similarity_openclip.py compare_image_similarity_openclip.py
 
 # Install system dependencies
 RUN apt-get update && apt-get install -y \
@@ -39,28 +39,12 @@ RUN apt-get update && apt-get install -y \
 
 RUN pip install --upgrade pip && pip install -r requirements.txt
 
-# Fontconfig Warning Suppression (optional)
-ENV FONTCONFIG_PATH=/etc/fonts
-ENV FONTCONFIG_FILE=/etc/fonts/fonts.conf
+# Create writable paths
+RUN mkdir -p /app/nltk_data /app/.config/matplotlib \
+ && mkdir -p /app/cache /app/data /app/OUTPUTS /app/outputs \
+ && mkdir -p /app/outputs/DETECTED_IMAGE /app/outputs/SCANNED_IMAGE /app/outputs/EXTRACTED_JSON \
+ && chmod -R 777 /app
 
-# Set writable path for nltk data
-ENV NLTK_DATA=/app/nltk_data
-RUN mkdir -p /app/nltk_data
-
-# Pre-download required NLTK models
-RUN python -m nltk.downloader -d /app/nltk_data \
-    punkt averaged_perceptron_tagger averaged_perceptron_tagger_eng
-
-RUN mkdir -p /app/cache /app/data && chmod -R 777 /app/cache /app/data
-RUN mkdir -p /app/OUTPUTS
-RUN mkdir -p /app/outputs
-
-RUN mkdir -p outputs/DETECTED_IMAGE outputs/SCANNED_IMAGE outputs/EXTRACTED_JSON
-
-# Expose the required port for HF Spaces
 EXPOSE 7860
 
-# Set the command to run your Flask app
-# CMD ["python", "extract_img_pdf.py"]
-# CMD ["python", "live_streaming_flask.py"]
 CMD ["python", "app_main.py"]