Update comments in Dockerfile (#26)

We could remove xtb and openmpi from the image,
but we keep them to speedup the app installation.

Dockerfile

@@ -4,14 +4,13 @@ LABEL maintainer="Daniel Hollas <[email protected]>"
 USER root
 WORKDIR /opt/
 
-# TODO: It looks like xtb-python is now available on Conda,
-# so we should just declare it with other dependencies in setup.cfg
-# https://pypi.org/project/xtb/#history
-# TODO: Move openmpi-installation to its own conda environment
-# as part of the aiidalab-ispg installation in `post_install` script,
-# see how QeApp does this.
 # NOTE: We bump traitlets package to 5.9 for performance and app loading speed.
 # This is a temporary measure until we update the underlying Jupyter image.
+# NOTE: We could remove the OpenMPI and xTB installations as we now can
+# install them directly during the aiidalab-ispg installation, see:
+# https://github.com/ispg-group/aiidalab-ispg/pull/221
+# but because it takes a non-trivial amount of time,
+# we install them here to speed up the installation.
 RUN mamba install --yes -c conda-forge \
      xtb-python \
      openmpi=4.1.1 \

README.md

@@ -4,9 +4,8 @@ Dockerfile for AiidaLab ATMOSPEC deployment.
 The following extra things are added on top of the `aiidalab/full-stack` image:
 
 1. SLURM queuing manager
-2. Additional conda packages (xtb-python, OpenMPI)
-   which currently must be installed in root conda environment.
-3. HTTPS support
+2. Some conda packages (e.g. OpenMPI for parallel ORCA)
+3. TODO: HTTPS support
 
 Original images taken from https://github.com/aiidalab/aiidalab-docker-stack
 

opt/setup-ispg-things.sh

@@ -1,9 +1,9 @@
-# Note: We defer the orca code installation for the post_install script
-# that should be run after aiidalab-ispg app is installed.
+# Here we setup a new AiiDA computer configured with SLURM.
+# Note: We defer the AiiDA ORCA code installation for the 'post_install' script
+# that is run when aiidalab-ispg app is installed.
 
 set -euo pipefail
 
-# However, we do setup a new computer configured with SLURM.
 computer_name=slurm
 
 # TODO: Use YAML for the config