[Project] Init Docker

Dockerfile

@@ -0,0 +1,153 @@
+### ------
+### The builder image, used to build the virtual environment
+### ------
+
+# Start with a minimal Python 3.11 base image based on Debian Buster
+FROM python:3.11-slim-buster
+
+### ------
+### Set up user
+### ------
+
+RUN useradd -m -u 1000 user
+
+### ------
+### Set up Python environment variables
+### ------
+
+# Ensures Python output is sent directly to the terminal (no buffering)
+ENV PYTHONUNBUFFERED=1 \
+    # Prevents Python from writing .pyc files (compiled bytecode)
+    PYTHONDONTWRITEBYTECODE=1 \
+    ### pip configuration to optimize the installation process
+    # Do not cache pip packages
+    PIP_NO_CACHE_DIR=off \
+    # Disable pip version checking
+    PIP_DISABLE_PIP_VERSION_CHECK=on \
+    # Set a longer timeout for pip commands (100 seconds)
+    PIP_DEFAULT_TIMEOUT=100 \
+    ### poetry configuration to install and manage dependencies
+    # Set the version of Poetry to use
+    POETRY_VERSION=1.4.2 \
+    # Define the location where Poetry will be installed
+    POETRY_HOME="/opt/poetry" \
+    #POETRY_HOME="/home/user/.local" \
+    # Create the virtual environment inside the project folder
+    POETRY_VIRTUALENVS_IN_PROJECT=true \
+    # Ensure Poetry creates virtual environments for the project
+    POETRY_VIRTUALENVS_CREATE=1 \
+    # Prevent Poetry from asking for interactive input during install
+    POETRY_NO_INTERACTION=1 \
+    ### Define paths for the virtual environment and Python setup
+    # Define where Python setup files will reside
+    # Define where the virtual environment will be created
+    PYSETUP_PATH="/opt/pysetup" \
+    #PYSETUP_PATH="/opt/pysetup" \
+    VENV_PATH="/opt/pysetup/.venv" \
+   # PATH="$POETRY_HOME/bin:$VENV_PATH/bin:$PATH" \
+    ### Hugging Face
+    # Set Hugging Face to offline mode to avoid network access
+    HF_HUB_OFFLINE=0 \
+    NLTK_DATA="/home/user/nltk/" \
+    # Store files
+    DATA_DIR="/app/data/" \
+    # Set the custom Poetry configuration directory
+    POETRY_CONFIG_DIR="$POETRY_HOME/.config/pypoetry"  \
+    PASSLIB_BUILTIN_BCRYPT="enabled"
+
+# Add Poetry and virtual environment to the system PATH (Note add after declared ENV)
+ENV PATH="$POETRY_HOME/bin:$VENV_PATH/bin:$PATH"
+
+### ------
+### Package
+### ------
+
+# Update package lists and install required build dependencies
+RUN apt-get update \
+    && apt-get install --no-install-recommends -y \
+        # Install curl to download files from the internet
+        curl \
+        # Install essential build tools (for building Poetry and other packages)
+        build-essential \
+        wget \
+        git \
+        openssh-client \
+    && curl -sSL https://install.python-poetry.org | python3 - \
+    # Remove build tools after installation to reduce image size
+    && apt-get purge -y --auto-remove build-essential \
+    # Clean up apt cache to minimize image size
+    && apt-get clean \
+    # Remove leftover lists to further reduce image size
+    && rm -rf /var/lib/apt/lists/*
+
+### ------
+### Create DIRs
+### ------
+
+WORKDIR /app
+WORKDIR /repo
+
+### ------
+### Git Clone secret
+### ------
+
+RUN --mount=type=secret,id=TAG,mode=0444,required=true \
+    --mount=type=secret,id=URL,mode=0444,required=true \
+    git clone --branch $(cat /run/secrets/TAG) $(cat /run/secrets/URL) /repo
+
+### ------
+### Copy file
+### ------
+
+RUN cp -r /repo/api_writer/* /app/
+RUN cp -r /repo/gary /app/gary
+RUN rm -rf /repo
+
+### ------
+### Modif path of Lib
+### ------
+
+WORKDIR /app
+
+RUN sed -i 's/\.\.\/gary/gary/g' pyproject.toml
+
+### ------
+### Install project dependencies using Poetry
+### ------
+
+# Move to gary and install need optional lib
+WORKDIR /app/gary
+# Install only the dependencies needed (Mango and pymupdf)
+RUN poetry install --no-interaction --with mangodb --with pymupdf \
+    # Clean up the Poetry cache to reduce image size
+    && rm -rf $POETRY_CACHE_DIR
+
+WORKDIR /app
+
+# Optimize Poetry's performance with more parallel workers
+RUN poetry config installer.max-workers 10 \
+    # Install only the dependencies needed
+    && poetry install --no-interaction --no-dev \
+    # Clean up the Poetry cache to reduce image size
+    && rm -rf $POETRY_CACHE_DIR
+
+# Create the custom Poetry configuration directory and set permissions
+RUN mkdir -p $POETRY_CONFIG_DIR && \
+    # Create the NTLK folder (llama)
+    mkdir -p $NLTK_DATA && \
+    # Create the folder store files
+    mkdir -p $DATA_DIR && \
+    chown -R user:user $POETRY_CONFIG_DIR && \
+    chown -R user:user $NLTK_DATA && \
+    chown -R user:user $DATA_DIR  && \
+    chmod -R 777 $DATA_DIR
+
+### ------
+### Command to run
+### ------
+
+# Expose the port FastAPI will run on
+EXPOSE 7860
+
+# This command starts the API using Poetry's virtual environment
+CMD ["poetry", "run", "api"]