Bringing in image directory from deployment repo

Dockerfile

@@ -0,0 +1,139 @@
+FROM buildpack-deps:jammy-scm as base
+
+# Set up common env variables
+ENV TZ=America/Los_Angeles
+RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone
+
+ENV LC_ALL en_US.UTF-8
+ENV LANG en_US.UTF-8
+ENV LANGUAGE en_US.UTF-8
+ENV DEBIAN_FRONTEND=noninteractive
+ENV NB_USER jovyan
+ENV NB_UID 1000
+# These are used by the python, R, and final stages
+ENV CONDA_DIR /srv/conda
+ENV R_LIBS_USER /srv/r
+
+RUN apt-get -qq update --yes && \
+    apt-get -qq install --yes locales && \
+    echo "en_US.UTF-8 UTF-8" > /etc/locale.gen && \
+    locale-gen
+
+RUN adduser --disabled-password --gecos "Default Jupyter user" ${NB_USER}
+
+# Install all apt packages
+COPY apt.txt /tmp/apt.txt
+RUN apt-get -qq update --yes && \
+    apt-get -qq install --yes --no-install-recommends \
+        $(grep -v ^# /tmp/apt.txt) && \
+    apt-get -qq purge && \
+    apt-get -qq clean && \
+    rm -rf /var/lib/apt/lists/*
+
+# Install R.
+# These packages must be installed into the base stage since they are in system
+# paths rather than /srv.
+# Pre-built R packages from rspm are built against system libs in jammy.
+ENV R_VERSION=4.4.1-1.2204.0
+ENV LITTLER_VERSION=0.3.19-1.2204.0
+RUN apt-key adv --keyserver keyserver.ubuntu.com --recv-keys E298A3A825C0D65DFD57CBB651716619E084DAB9
+RUN echo "deb https://cloud.r-project.org/bin/linux/ubuntu jammy-cran40/" > /etc/apt/sources.list.d/cran.list
+RUN curl --silent --location --fail https://cloud.r-project.org/bin/linux/ubuntu/marutter_pubkey.asc > /etc/apt/trusted.gpg.d/cran_ubuntu_key.asc
+RUN apt-get update -qq --yes > /dev/null && \
+    apt-get install --yes -qq \
+        r-base-core=${R_VERSION} \
+        r-base-dev=${R_VERSION} \
+        littler=${LITTLER_VERSION} > /dev/null
+
+ENV RSTUDIO_URL=https://download2.rstudio.org/server/jammy/amd64/rstudio-server-2024.04.2-764-amd64.deb
+RUN curl --silent --location --fail ${RSTUDIO_URL} > /tmp/rstudio.deb && \
+    apt install --no-install-recommends --yes /tmp/rstudio.deb && \
+    rm /tmp/rstudio.deb
+
+ENV SHINY_SERVER_URL https://download3.rstudio.org/ubuntu-18.04/x86_64/shiny-server-1.5.20.1002-amd64.deb
+RUN curl --silent --location --fail ${SHINY_SERVER_URL} > /tmp/shiny.deb && \
+    apt install --no-install-recommends --yes /tmp/shiny.deb && \
+    rm /tmp/shiny.deb
+
+# Set CRAN mirror to rspm before we install anything
+COPY Rprofile.site /usr/lib/R/etc/Rprofile.site
+# RStudio needs its own config
+COPY rsession.conf /etc/rstudio/rsession.conf
+
+# R_LIBS_USER is set by default in /etc/R/Renviron, which RStudio loads.
+# We uncomment the default, and set what we wanna - so it picks up
+# the packages we install. Without this, RStudio doesn't see the packages
+# that R does.
+# Stolen from https://github.com/jupyterhub/repo2docker/blob/6a07a48b2df48168685bb0f993d2a12bd86e23bf/repo2docker/buildpacks/r.py
+# To try fight https://community.rstudio.com/t/timedatectl-had-status-1/72060,
+# which shows up sometimes when trying to install packages that want the TZ
+# timedatectl expects systemd running, which isn't true in our containers
+RUN sed -i -e '/^R_LIBS_USER=/s/^/#/' /etc/R/Renviron && \
+    echo "R_LIBS_USER=${R_LIBS_USER}" >> /etc/R/Renviron && \
+    echo "TZ=${TZ}" >> /etc/R/Renviron
+
+# =============================================================================
+# This stage exists to build /srv/r.
+FROM base as srv-r
+
+# Create user owned R libs dir
+# This lets users temporarily install packages
+RUN install -d -o ${NB_USER} -g ${NB_USER} ${R_LIBS_USER}
+
+# Install R libraries as our user
+USER ${NB_USER}
+
+COPY class-libs.R /tmp/class-libs.R
+RUN mkdir -p /tmp/r-packages
+
+# Our install.R needs devtools which needs install2.r which needs docopt.
+# install2.r is not reproducible, but our install.R script is.
+RUN Rscript -e "install.packages('docopt')"
+RUN /usr/lib/R/site-library/littler/examples/install2.r devtools
+
+# Install all our base R packages
+COPY install.R  /tmp/install.R
+RUN /tmp/install.R && rm -rf /tmp/downloaded_packages
+
+# =============================================================================
+# This stage exists to build /srv/conda.
+FROM base as srv-conda
+
+COPY install-mambaforge.bash /tmp/install-mambaforge.bash
+RUN /tmp/install-mambaforge.bash
+
+# Install conda environment as our user
+USER ${NB_USER}
+
+ENV PATH ${CONDA_DIR}/bin:$PATH
+
+COPY infra-requirements.txt /tmp/infra-requirements.txt
+COPY environment.yml /tmp/environment.yml
+
+RUN mamba env update -p ${CONDA_DIR} -f /tmp/environment.yml && \
+    mamba clean -afy
+
+# =============================================================================
+# This stage consumes base and import /srv/r and /srv/conda.
+FROM base as final
+COPY --from=srv-r /srv/r /srv/r
+COPY --from=srv-conda /srv/conda /srv/conda
+
+# Install IR kernelspec. Requires python and R.
+ENV PATH ${CONDA_DIR}/bin:${PATH}:${R_LIBS_USER}/bin
+RUN R -e "IRkernel::installspec(user = FALSE, prefix='${CONDA_DIR}')"
+
+# clear out /tmp
+USER root
+#COPY postBuild /tmp/postBuild
+#RUN chmod +x /tmp/postBuild
+#RUN /tmp/postBuild
+RUN rm -rf /tmp/*
+
+
+USER ${NB_USER}
+WORKDIR /home/${NB_USER}
+
+EXPOSE 8888
+
+ENTRYPOINT ["tini", "--"]

Rprofile.site

@@ -0,0 +1,23 @@
+# Use RStudio's CRAN mirror to get binary packages. "latest" just means it has
+# all available versions. We can specify version numbers in
+# devtools::install_version.
+options(
+  repos = c(
+    CRAN = "https://packagemanager.rstudio.com/all/__linux__/jammy/latest"
+  )
+)
+
+# RStudio sets the HTTPUserAgent option which causes it to download binaries,
+# rather than source, from the Posit Public Package Manager. In order for the
+# RStudio terminal, Jupyter terminal, Jupyter R notebook, or anything else to
+# do the same, we must manually set HTTPUserAgent. This speeds up package
+# installation.
+# We fake rstudio version because `RStudio.Version()$version` is only available
+# when we are in RStudio.
+options(
+  HTTPUserAgent = sprintf(
+    "RStudio Server (%s); R (%s)",
+    "2024.04.2.764",
+    paste(getRversion(), R.version$platform, R.version$arch, R.version$os)
+  )
+)

apt.txt

@@ -0,0 +1,37 @@
+# our entry point
+tini
+
+# useful utilities
+tar
+less
+vim
+
+# for nbconvert, knitr, etc.
+texlive-xetex
+texlive-fonts-recommended
+texlive-plain-generic
+# provides FandolSong-Regular.otf for issue #2714
+texlive-lang-chinese
+
+# for notebook-as-pdf
+libx11-xcb1
+libxtst6
+libxrandr2
+libasound2
+libpangocairo-1.0-0
+libatk1.0-0
+libatk-bridge2.0-0
+libgtk-3-0
+libnss3
+libxss1
+fonts-noto-color-emoji
+
+# for rstudio
+#psmisc
+#sudo
+#libapparmor1
+lsb-release
+libclang-dev
+
+# Used by littler for IRkernel::installspec
+libzmq5

class-libs.R

@@ -0,0 +1,20 @@
+#!/usr/bin/env Rscript
+
+class_libs_install_version <- function(class_name, class_libs) {
+  print(paste("Installing packages for", class_name))
+  for (i in seq(1, length(class_libs), 2)) {
+    installed_packages <- rownames(installed.packages())
+    package_name = class_libs[i]
+    version = class_libs[i+1]
+    # Only install packages if they haven't already been installed!
+    # devtools doesn't do that by default
+    if (!package_name %in% installed_packages) {
+      print(paste("Installing", package_name, version))
+      devtools::install_version(package_name, version, quiet=TRUE)
+    } else {
+      # FIXME: This ignores version incompatibilities :'(
+      print(paste("Not installing", package_name, " as it is already installed"))
+    }
+  }
+  print(paste("Done installing packages for", class_name))
+}

environment.yml

@@ -0,0 +1,21 @@
+dependencies:
+- python=3.11.* # mamba installation with py3.11 fails: No module named 'conda.cliconda'
+- pip=23.1.*
+- nodejs=18.*
+# bug w/notebook and traitlets: https://github.com/jupyter/notebook/issues/7048
+- traitlets=5.9.*
+
+- jupyter-server-proxy==4.2.0
+- jupyter-rsession-proxy==2.2.0
+
+- syncthing==1.23.5
+- pyppeteer==1.0.2
+
+# for nbconvert
+- pandoc==3.1.3
+
+# for jupyter-tree-download
+- zip==3.0
+- pip:
+  - -r infra-requirements.txt
+  - jupyter-shiny-proxy==1.1

infra-requirements.txt

@@ -0,0 +1,29 @@
+# WARNING: Original source at scripts/infra-packages/requirements.txt
+# PLEASE DO NOT EDIT ELSEWHERE
+# After editing scripts/infra-packages/requirements.txt, please run
+# scripts/infra-packages/sync.bash.
+
+# This file pins versions of notebook related python packages we want
+# across all hubs. This makes sure we don't need to upgrade them
+# everwhere one by one.
+
+# FIXME: Freeze this to get exact versions of all dependencies
+notebook==7.0.7
+jupyterlab==4.0.11
+nbgitpuller==1.2.1
+jupyter-resource-usage==1.0.1
+# Matches version in images/hub/Dockerfile
+jupyterhub==4.1.6
+appmode==0.8.0
+ipywidgets==8.0.7
+jupyter-tree-download==1.0.1
+git-credential-helpers==0.2
+# Measure popularity of different packages in our hubs
+# https://discourse.jupyter.org/t/request-for-implementation-instrument-libraries-actively-used-by-users-on-a-jupyterhub/7994?u=yuvipanda
+git+https://github.com/shaneknapp/python-popularity-contest.git@add-error-handling
+# RISE is useful for presentations - see https://github.com/berkeley-dsep-infra/datahub/issues/2527
+RISE==5.7.1
+# syncthing for dropbox-like functionality
+jupyter-syncthing-proxy==1.0.3
+# jupyter archival tool for easy user downloads
+jupyter-archive==3.4.0

install-mambaforge.bash

@@ -0,0 +1,41 @@
+#!/bin/bash
+
+#
+# Download and install a pinned version of mambaforge.
+#
+
+set -ex
+
+cd $(dirname $0)
+MAMBAFORGE_VERSION=24.3.0-0
+
+URL="https://github.com/conda-forge/miniforge/releases/download/${MAMBAFORGE_VERSION}/Mambaforge-${MAMBAFORGE_VERSION}-Linux-x86_64.sh"
+INSTALLER_PATH=/tmp/mambaforge-installer.sh
+
+wget --quiet $URL -O ${INSTALLER_PATH}
+chmod +x ${INSTALLER_PATH}
+
+bash ${INSTALLER_PATH} -b -p ${CONDA_DIR}
+export PATH="${CONDA_DIR}/bin:$PATH"
+
+# Do not attempt to auto update conda or dependencies
+conda config --system --set auto_update_conda false
+conda config --system --set show_channel_urls true
+
+# Empty the conda history file, which seems to result in some effective pinning
+# of packages in the initial env, which we don't intend. This file must not be
+# removed.
+> ${CONDA_DIR}/conda-meta/history
+
+# Clean things out!
+conda clean --all -f -y
+
+# Remove the big installer so we don't increase docker image size too much
+rm ${INSTALLER_PATH}
+
+# Remove the pip cache created as part of installing mambaforge
+rm -rf ${HOME}/.cache
+
+chown -R $NB_USER:$NB_USER ${CONDA_DIR}
+
+conda list -n root

install.R

@@ -0,0 +1,18 @@
+#!/usr/bin/env r
+
+# Install devtools so we can install versioned packages
+#install.packages("devtools")
+
+source("/tmp/class-libs.R")
+
+# R packages to be installed that aren't from apt
+# Combination of informal requests & rocker image suggestions
+# Some of these were already in datahub image
+cran_packages = c(
+  "BiocManager", "1.30.21",
+  "IRkernel", "1.3.2",
+  "rmarkdown", "2.22",
+  "shiny", "1.7.4"
+)
+
+class_libs_install_version("Base packages", cran_packages)

rsession.conf

@@ -0,0 +1,2 @@
+# Use binary packages!
+r-cran-repos=https://packagemanager.rstudio.com/all/__linux__/jammy/latest