2022-11-07 jupyterlab surface topography docker recipe

docker/jupyterlab/README.md

@@ -0,0 +1,4 @@
+# Content
+
+Container recipes derived from the Jupyter Docker Stacks at https://jupyter-docker-stacks.readthedocs.io
+

docker/jupyterlab/jupyterlab-SurfaceTopography/Dockerfile

@@ -0,0 +1,63 @@
+# Build from within repository root 
+#
+# Modify conda enviornment running jupyter notebooks
+FROM jupyter/minimal-notebook:latest
+
+USER root
+
+COPY --chown=${NB_UID}:${NB_GID} jupyterlab-SurfaceTopography/conda-requirements.in "/home/${NB_USER}/"
+RUN pip install pip-tools
+RUN pip-compile "/home/${NB_USER}/conda-requirements.in" > "/home/${NB_USER}/conda-requirements.txt"
+RUN pip install --quiet --no-cache-dir --requirement "/home/${NB_USER}/conda-requirements.txt" && \
+    fix-permissions "${CONDA_DIR}" && \
+    fix-permissions "/home/${NB_USER}"
+
+# FROM imkteksim/dtool-jupyter:latest
+# Install from requirements.txt file
+RUN apt-get --yes update && \
+    apt-get --yes install \
+        clang cmake curl g++ gdb git m4 wget \
+        libboost-test-dev \ 
+        libcurl4-openssl-dev \
+        libeigen3-dev \
+        libfftw3-dev \
+        libgmp-dev \
+        libnetcdf-dev \
+        libopenblas-base \
+        libopenblas-dev \
+        python3-pip \
+        python3-dev \
+        python3-netcdf4 && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
+
+# Workaround for building packages and introducing an ipykernel outside of the jupyterlab conda environment
+RUN PATH=$(echo "$PATH" | sed -e 's|/opt/conda[^:]*:||g') && ( \
+        pip install --no-binary numpy SurfaceTopography && \
+        pip install ipykernel && \
+        python3 -m ipykernel install --name SurfaceTopography --display-name "Python (SurfaceTopography)") && \
+    fix-permissions "${CONDA_DIR}" && \
+    fix-permissions "/home/${NB_USER}"
+
+# Workaround for modifying system python outside of the jupyterlab conda environment
+COPY --chown=${NB_UID}:${NB_GID} jupyterlab-SurfaceTopography/requirements.txt "/home/${NB_USER}/"
+RUN PATH="$(echo "$PATH" | sed -e 's|/opt/conda[^:]*:||g')" && ( \
+        pip install --quiet --no-cache-dir --requirement "/home/${NB_USER}/requirements.txt" )
+
+COPY jupyterlab-SurfaceTopography/start-wrapper.sh /usr/local/bin/
+
+# Custom fonts config
+COPY jupyterlab-SurfaceTopography/etc/fonts.conf /etc/fonts/conf.avail/99-injected.conf
+RUN ln -s /etc/fonts/conf.avail/99-injected.conf /etc/fonts/conf.d/99-injected.conf
+RUN mkdir -p /fonts
+
+# Install from requirements.txt file
+RUN apt-get --yes update && \
+    apt-get --yes install fontconfig && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/*
+
+WORKDIR /home/${NB_USER}/work
+USER ${NB_UID}
+
+CMD ["start-wrapper.sh", "start-notebook.sh"]

docker/jupyterlab/jupyterlab-SurfaceTopography/README.md

@@ -0,0 +1,161 @@
+# Sample setup jupyterlab-SurfaceTopography
+
+## Changes to the container
+
+This docker image bases on the [Jupyter Docker Stacks](https://jupyter-docker-stacks.readthedocs.io/en/latest/).
+These base images use condato run a JupyterLab instance. Extensions to the Jupyter ecosystem (i.e. `jupytext`) 
+must hence go into the `conda-requirements.in` file. 
+
+`SurfaceTopography` does not work well within a conda environment. Instead, this image introduces a separate 
+ipython kernel `SurfaceTopography` that just uses the container-internal system Python and a few system libraries,
+i.e. `numpy`. Anything needed for production in connection with `SurfaceTopography` goes in `requirements.in` or
+into `requirements.txt` directly. When modifying `requirements.in`, regenerate `requirments.txt`with pinned versions
+as described below.
+
+If in need to quickly modify the `SurfaceTopography` kernel environment at runtime, pay attention to
+breaking out of the default conda enviornment in the shell
+
+* `conda deactivate`, and
+* removing any conda-related artifacts from the `$PATH`, i.e. with `export PATH=$(echo "$PATH" | sed -e 's|/opt/conda[^:]*:||g')`
+
+The latter is necessary to avoid conda compiler interfering with custom build steps when compiling against system libraries.
+
+## Python requirements
+
+To (re-)generate a `requirements.txt` with fixed package versions, use
+
+```console
+python3 -m venv venv
+source venv/bin/activate
+
+pip install --upgrade pip
+pip install wheel
+pip install pip-tools
+
+pip-compile requirements.in > requirements.txt
+```
+
+## Build container
+
+Build container from parent directory (`docker/jupyterlab`) with
+
+    docker build -t imteksim/jupyterlab-surfacetopography -f jupyterlab-SurfaceTopography/Dockerfile .
+
+or write directly to tar for transfer with
+
+    docker build --output type=tar,dest=imteksim-jupyterlab-SurfaceTopography.tar -t imteksim/jupyterlab-surfacetopography -f jupyterlab-SurfaceTopography/Dockerfile .
+
+and load at target destination with
+
+    docker load --input imteksim-jupyterlab-SurfaceTopography.tar
+
+## Certificates with acme.sh
+
+Identify the public hostname used for issuing SSL certificates, i.e.
+
+    export HOSTNAME="SOME-UUID.fr.bw-cloud-instance.org"
+
+for a public bw-cloud instance and a seecure directory to hold certificates, i.e.
+
+    export CERTDIR=${HOME}/acme.sh
+
+and use
+
+    mkdir -p acme.sh
+
+    # zerossl
+    docker run -v "${CERTDIR}:/acme.sh" --rm neilpang/acme.sh:latest \
+      acme.sh --register-account -m ${MY_EMAIL_ADDRESS}
+    docker run -v "${CERTDIR}:/acme.sh" -p 80:80 --rm neilpang/acme.sh:latest \
+      acme.sh --issue -d "${HOSTNAME}" --standalone --force
+
+    # or, letsencrypt
+    docker run -v "${CERTDIR}:/acme.sh" --rm neilpang/acme.sh:latest \
+      acme.sh --server letsencrypt --register-account -m ${MY_EMAIL_ADDRESS}
+    docker run -v "${CERTDIR}:/acme.sh" -p 80:80 --rm neilpang/acme.sh:latest \
+      acme.sh --server letsencrypt --issue -d "${HOSTNAME}" --standalone --force
+
+    sudo chown -R $USER:$USER $HOME/acme.sh
+
+to (re-)issue certiicates.
+
+## Fonts
+
+Microsoft fonts can be installed with
+
+    sudo apt-get install -y fontconfig
+    sudo apt-get install -y ttf-mscorefonts-installer
+    sudo fc-cache -f
+
+Check with
+
+    $ fc-match Arial
+    LiberationSans-Regular.ttf: "Liberation Sans" "Regular"
+
+Find location of installed fonts with
+
+    $ dpkg-query -L ttf-mscorefonts-installer
+    /.
+    /usr
+    /usr/lib
+    /usr/lib/msttcorefonts
+    /usr/lib/msttcorefonts/update-ms-fonts
+    /usr/share
+    /usr/share/doc
+    /usr/share/doc/ttf-mscorefonts-installer
+    /usr/share/doc/ttf-mscorefonts-installer/README.Debian
+    /usr/share/doc/ttf-mscorefonts-installer/changelog.gz
+    /usr/share/doc/ttf-mscorefonts-installer/copyright
+    /usr/share/fonts
+    /usr/share/fonts/truetype
+    /usr/share/lintian
+    /usr/share/lintian/overrides
+    /usr/share/lintian/overrides/ttf-mscorefonts-installer
+    /usr/share/package-data-downloads
+    /usr/share/package-data-downloads/ttf-mscorefonts-installer
+    /var
+    /var/lib
+    /var/lib/msttcorefonts
+
+## Run Jupyter Lab
+
+Identify a persistent workspace directory, i.e.
+
+    export WORKDIR=${HOME}/work
+
+and, optionally, a directory of additional fonts to make available to matplotlib, i.e. as above
+
+    export FONTDIR=/usr/share/fonts/truetype/msttcorefonts
+
+Run public Jupyter Lab on default https port with fixed password
+
+    docker run -d --rm -p 443:443 \
+        -v ${WORKDIR}:/home/jovyan/work \
+        -v ${FONTDIR}:/fonts \
+        -v ${CERTDIR}/${HOSTNAME}:/etc/ssl/notebook \
+        imteksim/jupyterlab-surfacetopography \
+        start-notebook.sh \
+        --ServerApp.keyfile=/etc/ssl/notebook/${HOSTNAME}.key \
+        --ServerApp.certfile=/etc/ssl/notebook/${HOSTNAME}.cer \
+        --ServerApp.password='argon2:$argon2id$v=19$m=10240,t=10,p=8$Jg1gFoiE0ybmBCVG+1s6uA$Jxz4/1591Z7so7JK2M1lRA' \
+        --ServerApp.ip='*' \
+        --ServerApp.port=443
+
+Note that the hash of the password is passed.
+
+    argon2:$argon2id$v=19$m=10240,t=10,p=8$Jg1gFoiE0ybmBCVG+1s6uA$Jxz4/1591Z7so7JK2M1lRA
+
+corresponds to password `imtek-simulation` and has been generated with
+
+    from notebook.auth import passwd
+    passwd()
+
+Add
+
+    --user root -e GRANT_SUDO=yes -e NB_GID=100 -e NB_USER=jovyan
+
+to the docker command line arguments if you need passwordless `sudo` within the container.
+
+## Access Jupyter Lab web interface
+
+Make sure `https` port 443 is reachable, navigate browser to `https://${HOSTNAME}` and log in with password specified on Jupyter Lab launch.

docker/jupyterlab/jupyterlab-SurfaceTopography/conda-requirements.in

@@ -0,0 +1,2 @@
+jupyterlab_autorun_cells
+jupytext

docker/jupyterlab/jupyterlab-SurfaceTopography/etc/fonts.conf

@@ -0,0 +1,6 @@
+<?xml version="1.0"?>
+<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
+<!-- ~/.fonts.conf for per-user font configuration -->
+<fontconfig>
+<dir>/fonts</dir>
+</fontconfig>

docker/jupyterlab/jupyterlab-SurfaceTopography/requirements.in

@@ -0,0 +1,36 @@
+click
+pep517
+tomli
+adhesion
+ase
+asgiref
+contactmechanics
+cycler
+dtool
+dtool_lookup_api
+dtoolcore
+dtool-s3
+gromacs
+GromacsWrapper
+imteksimfw
+ipython
+ipywidgets
+Jinja2
+matplotlib
+MDAnalysis
+miniball
+nglview
+numpy
+openpyxl
+ovito<=3.6.0
+pandas
+panedr
+ParmEd
+pymongo
+pytest
+PySide2
+PyYAML
+scipy
+scikit-learn
+seaborn
+git+https://github.com/IMTEK-Simulation/code-snippets.git