From 7e25aa5a46344f43102a256306e0905764f7b178 Mon Sep 17 00:00:00 2001
From: Adam Blake <adam@ablake.dev>
Date: Fri, 5 Jan 2024 13:01:56 -0800
Subject: [PATCH] refactor(datascience): simplify stages

---
 datascience-notebook/Dockerfile | 74 +++++++--------------------------
 1 file changed, 16 insertions(+), 58 deletions(-)

diff --git a/datascience-notebook/Dockerfile b/datascience-notebook/Dockerfile
index fb93889..0b3601c 100644
--- a/datascience-notebook/Dockerfile
+++ b/datascience-notebook/Dockerfile
@@ -22,12 +22,13 @@ ARG APT_LIB_DIR=/var/lib/apt
 ARG SHARED_LIBS_LIST=/tmp/shared-libs.txt
 
 
-# py-builder
+# base
 # -------------------------------------
-FROM ${PARENT_IMAGE} as py-builder
+FROM ${PARENT_IMAGE} as base
 
 USER root
 
+# install cmdstan dependencies
 ARG APT_CACHE_DIR APT_LIB_DIR
 # hadolint ignore=DL3008,DL3009
 RUN --mount=type=cache,target="${APT_CACHE_DIR}",sharing=locked \
@@ -35,15 +36,20 @@ RUN --mount=type=cache,target="${APT_CACHE_DIR}",sharing=locked \
   apt-get update && apt-get upgrade -y && apt-get -y install --no-install-recommends \
   g++
 
+USER ${NB_UID}
+
+# install cmdstan
+RUN Rscript -e 'remotes::install_cran("cmdstanr", repos = "https://mc-stan.org/r-packages/", dependencies = TRUE)' \
+  && Rscript -e 'cmdstanr::install_cmdstan()'
+
+
+# py-builder
+# -------------------------------------
+FROM base as py-builder
+
 RUN --mount=type=cache,target=${CONDA_DIR}/pkgs,uid=${NB_UID},sharing=locked \
   mamba install --yes \
-  "gdal" \
-  "pkg-config" \
-  "proj" \
-  && fix-permissions "${CONDA_DIR}" \
-  && fix-permissions "/home/${NB_USER}"
-
-USER ${NB_UID}
+  "gdal"
 
 # make wheels for the Python packages
 ARG PIP_CACHE_DIR PIP_WHEEL_DIR PY_REQUIREMENTS
@@ -53,59 +59,11 @@ RUN --mount=type=bind,source="${PY_REQUIREMENTS}",target=/tmp/requirements.txt \
   && pip3 wheel --cache-dir="${PIP_CACHE_DIR}" --wheel-dir="${PIP_WHEEL_DIR}" -r /tmp/requirements.txt
 
 
-# r-builder
-# -------------------------------------
-FROM ${PARENT_IMAGE} as r-builder
-
-
-# pre-shared-libs
-# -------------------------------------
-FROM ${BASE_IMAGE} as pre-shared-libs
-ARG SHARED_LIBS_LIST
-RUN find-shared-libs "${R_HOME}/library" > "${SHARED_LIBS_LIST}"
-
-
-# new-shared-libs
-# -------------------------------------
-FROM r-builder as new-shared-libs
-ARG SHARED_LIBS_LIST
-COPY --from=pre-shared-libs "${SHARED_LIBS_LIST}" /tmp/pre-shared-libs.txt
-RUN find-shared-libs "${R_HOME}/library" > /tmp/post-shared-libs.txt \
-  && diff-shared-libs /tmp/pre-shared-libs.txt /tmp/post-shared-libs.txt > "${SHARED_LIBS_LIST}"
-
-
 # final
 # -------------------------------------
-FROM ${BASE_IMAGE} as final
+FROM base
 LABEL maintainer="CourseKata <tech@coursekata.org>"
 
-USER root
-
-# install cmdstan dependencies
-ARG APT_CACHE_DIR APT_LIB_DIR
-# hadolint ignore=DL3008,DL3009
-RUN --mount=type=cache,target="${APT_CACHE_DIR}",sharing=locked \
-  --mount=type=cache,target="${APT_LIB_DIR}",sharing=locked \
-  apt-get update && apt-get upgrade -y && apt-get -y install --no-install-recommends \
-  g++
-
-# copy over R packages and shared libraries
-ARG SHARED_LIBS_LIST
-COPY --from=r-builder --chown=${NB_UID}:${NB_GID} "${R_HOME}/" "${R_HOME}"
-RUN --mount=type=cache,from=r-builder,source=/,target=/mounted-r-builder \
-  --mount=type=bind,from=new-shared-libs,source="${SHARED_LIBS_LIST}",target="${SHARED_LIBS_LIST}" \
-  copy-shared-libs "${SHARED_LIBS_LIST}" /mounted-r-builder / \
-  # this is hard to copy over from the r-builder image
-  && Rscript -e 'remotes::install_cran("V8", force = TRUE)' \
-  # install cmdstan
-  && Rscript -e 'remotes::install_cran("cmdstanr", repos = "https://mc-stan.org/r-packages/", dependencies = TRUE)' \
-  && Rscript -e 'cmdstanr::install_cmdstan()' \
-  # ensure user has full control of package and home dirs
-  && fix-permissions "${CONDA_DIR}" \
-  && fix-permissions "/home/${NB_USER}"
-
-USER ${NB_UID}
-
 # install Python packages from wheels
 ARG PIP_CACHE_DIR PIP_WHEEL_DIR PY_REQUIREMENTS
 RUN --mount=type=bind,source="${PY_REQUIREMENTS}",target=/tmp/requirements.txt \