# =============================================================================
# Databricks Container Services — minimal base image for client.6.x
#
# This Python-focused reference builds the essential Python environment used by
# the v6-standard image. It does not include the JDK or other non-Python runtime
# components. Add customer-specific packages only at the marked points.
# =============================================================================
FROM ubuntu:24.04

# Override this at build time when your organization uses a private PyPI mirror.
ARG PIP_INDEX_URL=https://pypi.org/simple
# Work around the Ubuntu 24.04 FIPS detection bug (LP#2066990).
ENV OPENSSL_FORCE_FIPS_MODE=0

WORKDIR /databricks

# Set the system timezone to UTC without persisting a runtime environment variable.
RUN ln -snf /usr/share/zoneinfo/Etc/UTC /etc/localtime \
    && echo "Etc/UTC" > /etc/timezone

# Keep the Linux package set aligned with the v6 client-image sources instead
# of pruning individual libraries based on assumed customer usage.
RUN apt-get update \
    && apt-get install -y \
        apt-transport-https \
        build-essential \
        ca-certificates \
        cmake \
        curl \
        file \
        git \
        iproute2 \
        iputils-ping \
        jq \
        libbz2-dev \
        libffi8 \
        libffi-dev \
        liblzma-dev \
        libpq-dev \
        libsqlite3-0 \
        libsqlite3-dev \
        libssl3t64 \
        libssl-dev \
        libtinfo6 \
        locales \
        procps \
        psmisc \
        python-is-python3 \
        python3.12 \
        python3.12-dev \
        sudo \
        unzip \
        wget \
        zstd \
    && apt-get clean \
    && rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*

# === ADD YOUR CUSTOM SYSTEM PACKAGES HERE ===
# Example: RUN apt-get update && apt-get install -y your-package && apt-get clean

RUN locale-gen en_US.UTF-8 \
    && update-locale LC_ALL=en_US.UTF-8 LANGUAGE=en_US.UTF-8
ENV LANG=en_US.UTF-8

# Databricks runs cluster-scoped library installations as this low-privilege user.
RUN useradd libraries

# Install uv system-wide using Astral's standalone installer.
RUN curl -LsSf https://astral.sh/uv/0.11.16/install.sh \
    | env UV_INSTALL_DIR=/usr/local/bin sh

# Seed a shared uv cache so the platform can create child virtual environments
# offline with the expected pip version.
RUN mkdir -p /opt/uv-cache \
    && uv --cache-dir /opt/uv-cache venv /tmp/uv-seed-venv \
    && UV_DEFAULT_INDEX="${PIP_INDEX_URL}" uv pip install --cache-dir /opt/uv-cache \
        --python /tmp/uv-seed-venv pip==25.3 \
    && rm -rf /tmp/uv-seed-venv

# Databricks requires the root virtual environment at this exact path.
RUN UV_DEFAULT_INDEX="${PIP_INDEX_URL}" uv venv --python=/usr/bin/python3.12 \
        --system-site-packages --seed --offline --cache-dir /opt/uv-cache \
        /databricks/python3

RUN mkdir -p /databricks/.virtualenv-def
COPY pyproject.toml /databricks/.virtualenv-def/pyproject.toml
RUN UV_DEFAULT_INDEX="${PIP_INDEX_URL}" uv pip install --no-cache \
        --python /databricks/python3/bin/python \
        --requirement /databricks/.virtualenv-def/pyproject.toml

# === ADD YOUR CUSTOM PYTHON PACKAGES HERE ===
# Example: RUN uv pip install --python /databricks/python3/bin/python your-package

# Register the PySpark and Delta libraries bundled with Databricks Connect in
# Python package metadata. Keep this after all Python package installations.
RUN /databricks/python3/bin/_databricks-connect-register-pyspark-installation

# ---------------------------------------------------------------------------
# Environment setup (required by Databricks platform — do not modify)
# ---------------------------------------------------------------------------
# The following files and symlink are required by Databricks.
RUN echo "client.6.x-base" > /databricks/IMAGE_KEY \
    && echo "client.6.2" > /databricks/DBR_VERSION \
    && ln -sfn /databricks/python3 /databricks/python

# /etc/environment is read by Databricks when the container starts.
# These variables are required by the platform; do not modify or remove them.
RUN printf '%s\n' \
    'CLUSTER_DB_HOME=/databricks' \
    'DATABRICKS_RUNTIME_VERSION=client.6.2' \
    'DB_HOME=/databricks' \
    'DEFAULT_DATABRICKS_ROOT_VIRTUALENV_ENV=/databricks/python3' \
    'IS_CLIENT_IMAGE=true' \
    'MLFLOW_DEPLOYMENTS_TARGET=databricks' \
    'MLFLOW_GATEWAY_URI=databricks' \
    'MLFLOW_PYTHON_EXECUTABLE=/databricks/python/bin/python' \
    'MLFLOW_TRACKING_URI=databricks' \
    'OPENSSL_FORCE_FIPS_MODE=0' \
    > /etc/environment

# === ADD YOUR CUSTOM ENV VARS HERE ===
# Example: RUN echo 'MY_VAR=my_value' >> /etc/environment