USGS ISIS3 Dockerfile

USGS Integrated Software for Imagers and Spectrometers v3 (ISIS3) Docker container.

This (docker) container recipes we provide here is based on the official ISIS install document (at Github). ISIS is installed with Conda in its own environment (isis) under /opt/conda).

The containers are based on osgeo/gdal images (by default, the gdal-ubuntu-small-*), though you can use the build-argument (--build-arg BASE_IMAGE=...) to specify any other (ubuntu) base image.

To not conflict with other common commands (e.g., pip), conda binary path is positioned at last in $PATH environment variables. This note is to mind you about paying attention when installing new packages: ISIS environment should be managed by conda while python packages should be managed by OS' pip/pip3.

The following ISIS3 environment variables:

ISISROOT=/opt/conda
ISISDATA=/isis/data
ISIS3DATA="${ISISDATA}"
ISISTESTDATA=/isis/data_test

Dockerfile.jupyter

The Jupyter-enabled container uses (by default) jupyter/base-notebook as base image (which, by default) uses ubuntu:focal as base image. If you want a jupyter/base-notebook based on osgeo/gdal you can do so by building a base-notebook image based on that just like we do for the other containers. Here is an example of how I do that.

ISIS3 documentation/resources:

How to run

If you are familiar with ISIS3 you know that ISIS needs its kernels/ancillary data -- the famous ISISDATA -- to do virtually anything. By default, in isis3 containers expect to find ISIS3DATA mounted at /isis/data directory.

Since "ISISDATA" is not included in the container the user must mount a local copy of "ISISDATA" as container's /isis/data. The examples below show how to do it.

ISIS3 command-line

In this example, the location for ISIS3DATA in the local/host filesystem is /path/to/isis3data. Running a isis3 container with the ancillary data (at /isis/data):

$ docker run -it --rm --name isis3 \
    -v "/path/to/isis3data":"/isis/data" \
    my_isis3_container

ISIS3 + GDAL command-line

Let's run the container with GDAL as in osgeo/gdal. ISIS is installed in its own conda environment, separated from the system.

Let's also bind (mount) our current directory ($PWD) to container's /mnt/data:

$ docker run -it --rm --name isis3_gdal \
    -v "$PWD":"/mnt/data" \
    -v "/path/to/isis3data":"/isis/data" \
    my_isis3_container:gdal

ISIS3 + Jupyter-Hub

If you would like a graphical interface to use ISIS, GDAL and many Python GIS libraries you should use the jupyterhub tag. It is based on chbrandt/gispy:gdal (which uses apt/pip as package managers) with Jupyter-Hub on top. ISIS is installed (as always) in its own conda environment.

$ docker run -it --rm --name isis3_jupyter \
    -p 8888:8888 \
    -v "$PWD":"/mnt/data" \
    -v "/path/to/isis3data":"/isis/data" \
    my_isis3_container:jupyterhub

Redefining ISIS3 environment variables

ISIS3 environment variables are defined in container's /etc/profile.d/isis.sh.

If you want to customize ISIS3 variables, overwrite the default bash-profile:

$ docker run -it --name isis3 \
    -v "$PWD/dockerfile/etc/isis.sh":"/etc/profile.d/isis.sh" \
    chbrandt/isis3

zhouxuanxiao/docker-isis3