Add `pyproject.toml` and `uv.lock` for dependency management configuration

Dockerfile

@@ -8,8 +8,21 @@ RUN apt-get update && apt-get install -y \
     git \
  && rm -rf /var/lib/apt/lists/*
 
-COPY requirements.txt ./
-RUN pip3 install --no-cache-dir -r requirements.txt
+# Install uv by copying from the official Docker image
+COPY --from=ghcr.io/astral-sh/uv:latest /uv /uvx /bin/
+
+# Copy the project metadata for dependency resolution
+COPY pyproject.toml uv.lock /app/
+
+# Install dependencies into a virtual environment
+# Use cache mount for better build performance
+RUN --mount=type=cache,target=/root/.cache/uv \
+    cd /app && \
+    uv sync --frozen --no-install-project
+
+# Set up the virtual environment
+ENV VIRTUAL_ENV=/app/.venv
+ENV PATH="/app/.venv/bin:$PATH"
 
 # Route all caches/configs to /app and set HOME for Streamlit and others
 ENV HOME=/app \
@@ -18,14 +31,14 @@ ENV HOME=/app \
     STREAMLIT_USER_PATH=/app/.streamlit \
     STREAMLIT_DISABLE_TELEMETRY=1 \
     PYTHONUNBUFFERED=1 \
-    PIP_NO_CACHE_DIR=1 \
+    PYTHONDONTWRITEBYTECODE=1 \
+    UV_COMPILE_BYTECODE=1 \
+    UV_LINK_MODE=copy \
     HF_HOME=/app/.cache/huggingface \
     TRANSFORMERS_CACHE=/app/.cache/huggingface/transformers \
     TORCH_HOME=/app/.cache/torch \
-    MPLCONFIGDIR=/app/.config/matplotlib \
     XDG_CONFIG_HOME=/app/.config \
     XDG_CACHE_HOME=/app/.cache \
-    # EasyOCR cache dir (Docling may trigger EasyOCR)
     EASY_OCR_CACHE_DIR=/app/.EasyOCR
 
 # Create writable dirs for all caches/configs, including EasyOCR
@@ -33,7 +46,6 @@ RUN mkdir -p \
     /app/.streamlit/cache \
     /app/.cache/huggingface/transformers \
     /app/.cache/torch \
-    /app/.config/matplotlib \
     /app/.config \
     /app/.cache \
     /app/.EasyOCR \

pyproject.toml

@@ -0,0 +1,14 @@
+[project]
+name = "document-hf"
+version = "0.1.0"
+description = "Streamlit app to compare Marker vs Docling PDF extraction"
+readme = "README.md"
+# Python requirement based on docling dependency constraints
+requires-python = ">=3.13"
+dependencies = [
+  "docling",
+  "marker-pdf",
+  "streamlit",
+  "st-diff-viewer",
+]
+