ISIS

Table of Contents

• Requests for Comment
• FAQ
• ISIS Tutorials
• Citing ISIS
• Installation
• Start Contributing
• ISIS Data Area
• ISIS Test Data
• Installing Older Versions of ISIS
• Semantic Versioning and It's Role in Describing the Software

Requests for Comment

The ISIS project uses a Request for Comment (RFC) model where major changes to the code base, data area, or binary delivery process are proposed, iterated on, and potentially adopted. Right now, RFCs are being housed in this repository's wiki.

Current open RFCs:

• No Requests for Comment are currently open

  We encourage all contributors and users to review open RFCs and comment, as these proposed changes will impact use of the software.

FAQ

We maintain a list of frequently encountered questions and issues. Before opening a new issue, please take a look at the FAQ.

ISIS Tutorials

Please refer to the GitHub wiki page ISIS Online Workshops for current ISIS tutorials.

Citing ISIS

The badge at the top of this README lists the DOI of the most recent ISIS version. As of 05/09/2024, the latest release of ISIS is version 8.0.3, and its DOI is 10.5066/P13YBMZA.

The Releases Page on GitHub lists the DOI for each version of ISIS. Older versions may be listed on Zenodo. It is good practice to cite the version of the software being used by the citing work, so others can reproduce your exact results.

Installation

This installation guide is for ISIS users interested in installing ISIS (3.6.0)+ through conda.

ISIS Installation With Conda

1. Download either the Anaconda or Miniconda installation script for your OS platform. Anaconda is a much larger distribtion of packages supporting scientific python, while Miniconda is a minimal installation and not as large: Anaconda installer, Miniconda installer

2. If you are running on some variant of Linux, open a terminal window in the directory where you downloaded the script, and run the following commands. In this example, we chose to do a full install of Anaconda, and our OS is Linux-based. Your file name may be different depending on your environment.

   chmod +x Anaconda3-5.3.0-Linux-x86_64.sh
   ./Anaconda3-5.3.0-Linux-x86_64.sh

   This will start the Anaconda installer which will guide you through the installation process.

3. If you are running Mac OS X, a pkg file (which looks similar to Anaconda3-5.3.0-MacOSX-x86_64.pkg) will be downloaded. Double-click on the file to start the installation process.

4. After the installation has finished, open up a bash prompt in your terminal window.

5. If you have an ARM64 Mac (M1/M2) running Catalina (or later), additional prerequisites must be installed for ISIS to run in emulation:

   • Install XQuartz. (Tested with XQuartz 2.8.5 on MacOS Catalina)
   • Install Rosetta2. From the terminal run: /usr/sbin/softwareupdate --install-rosetta --agree-to-license
   • Include the # MacOS ARM64 Only lines below

6. Next setup your Anaconda environment for ISIS. In the bash prompt, run the following commands:

   [!WARNING] ISIS 8.1.0 is incompatible with Python 3.10, 3.11, and 3.12 The conda create command below creates a conda environment with Python 3.9

   
   #MacOS ARM64 Only - Setup the new environment as an x86_64 environment
   export CONDA_SUBDIR=osx-64
   
   #Create a new conda environment to install ISIS in
   conda create -n isis python=3.9
   
   #Activate the environment
   conda activate isis
   
   #MacOS ARM64 Only - Force installation of x86_64 packages instead of ARM64
   conda config --env --set subdir osx-64
   
   #Add the following channels to the environment
   conda config --env --add channels conda-forge
   conda config --env --add channels usgs-astrogeology
   
   #Verify you have the correct channels:
   conda config --show channels
   
   #You should see:
   
   channels:
       - usgs-astrogeology
       - conda-forge
       - defaults
   
   #The order is important.  If conda-forge is before usgs-astrogeology, you will need to run:
   
   conda config --env --add channels usgs-astrogeology
   
   #Then set channel_priority to flexible in case there is a global channel_priority=strict setting
   conda config --env --set channel_priority flexible

7. The environment is now ready to download ISIS and its dependencies:

   conda install -c usgs-astrogeology isis

   [!NOTE] The install may take 1 to 2 hours.

   If you would like to download an LTS version, follow the following format below:

   conda install -c "usgs-astrogeology/label/LTS" isis=8.0.1

8. Finally, setup the environment variables:

   ISIS requires several environment variables to be set in order to run correctly. The variables include: ISISROOT and ISISDATA.

   More information about the ISISDATA environment variable and the ISIS Data Area can be found here.

   The following steps are only valid for versions of ISIS after 4.2.0. For older versions of ISIS follow the instructions in this readme file.

   There are two methods to configure the environment variables for ISIS:

   1. Using conda env config vars preferred

      Conda has a built in method for configuring environment variables that are specific to a conda environment since version 4.8. This version number applies only to the conda package, not to the version of miniconda or anaconda that was installed.

      To determine if your version of conda is recent enough run:

      conda --version

      If the version number is less than 4.8, update conda to a newer version by running:

      conda update -n base conda

      The version number should now be greater than 4.8.

      To use the built in environment variable configuration feature, first activate the environment by first running:

      conda activate isis

      After activation, the environment variables can be set using the syntax: conda config vars set KEY=VALUE. To set all the environment variables ISIS requires, run the following command, updating the path to ISISDATA as needed:

      conda env config vars set ISISROOT=$CONDA_PREFIX ISISDATA=[path to data directory]

      To make these changes take effect, re-activate the isis environment by running:

      conda activate isis

      The environment variables are now set and ISIS is ready for use every time the isis environment is activated.

      Note This method will not enable tab completion for arguments in C-Shell.

   2. Using the provided isisVarInit.py script:

      To use the default values for: $ISISROOT and $ISISDATA, run the ISIS variable initialization script with default arguments:

      python $CONDA_PREFIX/scripts/isisVarInit.py

      Executing this script with no arguments will result in $ISISROOT=$CONDA_PREFIX and $ISISDATA=$CONDA_PREFIX/data. The user can specify different directories for $ISISDATA using the optional value:

      python $CONDA_PREFIX/scripts/isisVarInit.py --data-dir=[path to data directory]

      Now every time the isis environment is activated, $ISISROOT and $ISISDATA will be set to the values passed to isisVarInit.py. This does not happen retroactively, so re-activate the isis environment with one of the following commands:

      for Anaconda 3.4 and up - conda activate isis prior to Anaconda 3.4 - source activate isis

Installation with Docker

The ISIS production Dockerfile automates the conda installation process above. You can either build the Dockerfile yourself or use the usgsastro/isis image from DockerHub.

To build the Dockerfile

1. Download the production Docker file
2. Build the Dockerfile

   docker build -t isis -f production.dockerfile .

3. Run the Dockerfile

   docker run -it isis bash

Run run the prebuilt image

docker run -it usgsastro/isis bash

Usage with the ISIS data area

Usually you'll want to mount an external directory containing the ISIS data. The data is not included in the Docker image.

docker run -v /my/data/dir:/opt/conda/data -v /my/testdata/dir:/opt/conda/testData -it usgsastro/isis bash

Then download the data into /my/data/dir to make it accessible inside your container.

Practical Usage with other conda packages

If you don't use conda for anything else on your computer, you can skip this section.

If you use conda to install other packages, you may run into difficulties with adding the isis conda package to those environments or adding other conda packages to the isis environment you just created above. This is because the isis conda package pins a number of requirements that may clash with other packages.

At this time, we recommend creating the isis environment as detailed above, and then not adding any other conda packages to it. This is similar to the best practice usage of not adding any conda packages to your 'base' conda environment.

Instead, when you need to have a conda environment with other packages that also needs to be able to run ISIS programs, we have two different options. In both cases, we'll assume that you create a new environment called 'working' (but it could be named anything) that you want to add some conda packages to, but from which you also want ISIS access.

The first step is to create 'working' and add whatever conda packages you want.

Easy mode, with stacking

1. conda activate isis

2. conda activate --stack working

That's it. Told you it was easy.

This activates the isis environment, gets it all set up, and then it 'stacks' the new working environment on top of it. To get out, you'll have to conda deactivate two times to get out of working and then out of isis.

Harder mode, with activation script hacking

The above stacking situation may have issues if you have a particularly complicated set of packages or other dependencies. The idea here is that the only thing you really need in your 'working' environment are the ISIS environment variables and the path to the ISIS executables.

If the above paragraph sounded like gibberish, please seek help from your system administrator or local computer guru.

And we can do this via customizations in the conda environment's activate.d/ and deactivate.d/ directories. Adding these things can also be done manually from the command line, but encoding them in the activate.d/ and deactivate.d/ scripts is handy.

1. Create your conda environment however you like, adding whatever packages you need. If you were reading the directions above, you've already done this.

2. Locate the path to your conda environments:

   conda activate
   echo $CONDA_PREFIX
   conda deactivate

   You'll probably get a directory that is in your home directory and is named anaconda3 or miniconda3 or something similar. For the rest of this set of instructions, we'll refer to it as $HOME/anaconda3 to represent a directory named anaconda3 in your home directory, but this should be whatever you get from the above echo command.

3. Locate the path to your ISIS conda environment:

   conda activate isis
   echo $CONDA_PREFIX
   conda deactivate

   This should probably be $HOME/anaconda3/envs/isis. You can see that it starts with whatever you got from step 1, and ends in the name of your isis environment, if you followed the installation instructions above, you called that environment 'isis'.

   You can do the same thing to find the path to your new 'working' environment, but in this example, it will be at $HOME/anaconda3/envs/working.

4. Copy the ISIS activation and deactivation scripts to your new environment. Please note that the directory names in the instructions below are based on how you installed conda and what you named the 'isis' environment and the 'working' environment. You may not be able to just copy and paste these instructions directly, they are an example. Likewise, if your shell doesn't take the bash syntax in the .sh files, then you may need to select one of the other env_vars.* files in the isis directories.

   cd $HOME/anaconda3/envs/
   mkdir -p working/etc/conda/activate.d/
   mkdir -p working/etc/conda/deactivate.d/
   cp isis/etc/conda/activate.d/env_vars.sh working/etc/conda/activate.d/env_vars.sh
   cp isis/etc/conda/deactivate.d/env_vars.sh working/etc/conda/deactivate.d/env_vars.sh

5. Edit the copied activation file in $HOME/anaconda3/envs/working/etc/conda/activate.d/ to add the ISIS executable directory to the path, by adding this line at the end:

   export PATH=$PATH:$ISISROOT/bin

   Or whatever is appropriate for your shell if you aren't using the .sh file. No matter how you do it, it is important that you add $ISISROOT/bin to the end of the current path in your working environment, and not at the beginning.

6. Edit the copied deactivation file in $HOME/anaconda3/envs/working/etc/conda/deactivate.d/ to remove the path, by adding this line at the end:

   export PATH=`echo -n $PATH | awk -v RS=: -v ORS=: '/isis/ {next} {print}' | sed 's/:$//'`;`

   Or whatever is appropriate for your shell if you aren't using the .sh file. If your ISIS environment is not called isis, then you need to replace that part in the awk line above. You can look in the activate.d/env_vars.sh file to see what it should be.

Adding the lines in steps 5 and 6 manually adds the 'bin/' directory of the ISIS environment to your path (step 5), and then manually removes it (step 6) on deactivation. If you are using some other shell, you may need to use a different syntax to add and remove these elements to and from your path.

Updating

To update to the newest version of ISIS, run conda update -c usgs-astrogeology isis

To update to our latest release candidate , run conda update -c usgs-astrogeology/label/RC isis

Note that for ISIS versions 3.10 and above, new versions and release candidates will only be available under the package name isis and conda update isis3 and conda update -c usgs-astrogeology -c usgs-astrogeology/label/RC isis3 will not work for additional updates. Instead, after installing an isis package, conda update isis should be used to update to a new version and conda update -c usgs-astrogeology/label/RC isis to update to a new release candidate.

Operating System Requirements

ISIS runs on many UNIX variants. ISIS does not run natively on MS Windows, although it has been successfully run on Windows 10 using the Windows Subsystem for Linux (WSL). Instructions for doing this can be found here. The UNIX variants ISIS has been successfully built on are:

• Ubuntu 18.04 LTS
• Mac OS X 10.13.6 High Sierra
• Fedora 28
• CentOS 7.2

ISIS may be run on other Linux or macOS operating systems then those listed above, but it has not been tested and is not supported.

Hardware Requirements

Here are the minimum hardware requirements

• 64-bit (x86) processors
• 2 GB RAM
• 2.5 GB of disk space for ISIS binaries
• 10 GB to 510 GB disk space for ISIS data
• 10 GB to many TB disk space for processing images
• A quality graphics card

To build and compile ISIS requires following the instructions listed below, which are given on the GitHub wiki page for the ISIS project:

• Getting Started With GitHub
• Building ISIS With cmake
• New ISIS environmental variables and their meanings
• Custom data and test directories
• Cleaning builds
• Building individual applications/objects
• Building ISIS documentation
• What to do if you encounter any problems

The ISIS Data Area

Ancillary Data

Many ISIS applications require ancillary data. For example, calibration applications require flat files to do flat field corrections, and map projection applications require DTMs to accurately compute intersections. Due to its size, this data is stored in a separate directory called the ISIS Data Area. Any location can be used for the ISIS Data Area, the software simply requires that the ISISDATA environment variable is set to its location.

Structure of the ISIS Data Area

Under the root directory of the ISIS Data Area pointed to by the ISISDATA/ISIS3DATA environment variable are a variety of sub-directories. Each mission supported by ISIS has a sub-directory that contains mission specific processing data such as flat files and mission specific SPICE. There are also data areas used by more generic applications. These sub-directories contain everything from templates to test data.

Versions of the ISIS Data Area

In ISIS version 4.1.0 and later, several files previously stored in the data area closely associated with ISIS applications were moved into version control with the ISIS source code. To support the use of data in ISIS versions predating 4.1.0 the downloadIsisData application will need to download the data named legacybase. This is explained further in the Full ISIS Data Download section.

Size of the ISIS Data Area

If you plan to work with data from all missions, then the download will require about 520 GB for all the ancillary data. However, most of this volume is taken up by SPICE files. We have a Web service that can be used in lieu of downloading all of the SPICE files. This reduces the total download size to about 10 GB.

Full ISIS Data Download

Warning: if you are looking to download ISIS data via rsync, this is no longer supported. The rsync server isisdist.astrogeology.usgs.gov was shutdown in November 30, 2022 and replaced with an Amazon S3 storage bucket specified in rclone.conf. The outdated rsync download information can be found here and updated instructions for downloading ISIS data are provided below.

The ISIS Data Area is hosted on a combination of AWS S3 buckets and public http servers e.g. NAIF, Jaxa, ESA and not through conda channels like the ISIS binaries. This requires using the downloadIsisData script from within a terminal window within your Unix distribution, or from within WSL if running Windows 10. Downloading all mission data requires over 520 GB of disk space. If you want to acquire only certain mission data click here. To download all ISIS data files, continue reading.

To download all ISIS data, use the following command:

downloadIsisData all $ISISDATA

Note: this applicaion takes in 3 parameters in the following order \ \ \
For more usage, run downloadIsisData --help or downloadIsisData -h.

Note: The above command downloads all ISIS data including the required base data area and all of the optional mission data areas.

Partial Download of ISIS Base Data

This data area contains data that is common between multiple missions such as DEMS and leap second kernels. As of ISIS 4.1, the base data area is no longer required to run many applications as data such as icons and templates has been moved into the binary distribution. If you plan to work with any applications that use camera models (e.g., cam2map, campt, qview), it is still recommended you download the base data area. To download the base data area run the following commands:

downloadIsisData base $ISISDATA

Note: For accessing ISIS Data for versions of ISIS prior to ISIS 4.1.0, you must download the legacybase area and not the base area when using this application as shown below:

downloadIsisData legacybase $ISISDATA

Partial Download of Mission Specific Data

There are many missions supported by ISIS. If you are only working with a few missions then you can save disk space by downloading only those specific data areas. If you want to limit the download even further, read the next section about the SPICE Web Service. Otherwise jump to the mission specific sections.

ISIS SPICE Web Service

ISIS can now use a service to retrieve the SPICE data for all instruments ISIS supports via the internet. To use this service instead of your local SPICE data, click the WEB check box in the spiceinit program GUI or type spiceinit web=yes at the command line. Using the ISIS SPICE Web Service will significantly reduce the size of the downloads from our data area. If you want to use this new service, without having to download all the SPICE data, add the following argument to the mission-specific downloadIsisData command:

--exclude="kernels/**"

For example:

downloadIsisData cassini $ISISDATA --exclude="kernels/**"

You can also use include argument to partially download specific kernels. For example, download only cks and fks of LRO mission:

downloadIsisData lro $ISISDATA --include="{ck/**,fk/**}"

WARNING: Some instruments require mission data to be present for radiometric calibration, which is not supported by the SPICE Web Server, and some programs that are designed to run an image from ingestion through the mapping phase do not have an option to use the SPICE Web Service. For information specific to an instrument, see the documentation for radiometric calibration programs.

Mission Specific Data Downloads

For versions of ISIS prior to ISIS 4.1.0, please use the --legacy flag

ISIS Test Data

ISIS is comprised of two types of tests, custom Makefile based tests, and GTest based tests. Those that are GTest based, make economical use of data that exists on the ISIS3 repo along with the source, so no special data is required to run those other than the ISIS data area. The Makefile tests depend on a separate source of data that consists of a few gigabytes of input and expected output data used for testing ISIS applications. The Makefile based tests use the ISISTESTDATA environment variable to know where the required data are located. The total size of this test data decreases as we work towards converting Makefile tests to GTests.

How to download the ISIS test data with rclone

Test data is hosted using Amazon S3 storage buckets. We recommend using rclone to pull the data into a local directory. You can download rclone using their instructions (see: https://rclone.org/downloads/) or by using an anaconda environment (see: https://docs.anaconda.com/anaconda/install/). If you already have an anaconda environment up, install rclone with: conda install –c conda-forge rclone

Once rclone is installed, with $ISISROOT set, simply run: rclone --config $ISISROOT/etc/isis/rclone.conf sync asc_s3:asc-isisdata/isis_testData/ $ISISTESTDATA where:

• $ISISTESTDATA is the environment variable defining the location of the isis test data.
• --config overwrites the default config path, you want to use the rclone config that ships with ISIS.
• asc_s3: is the name of S3 configuration in the configuration file that ships with ISIS. This can be whatever you want to name it, in this case it is named remote.
• asc-isisdata/isis_testData/ is the name of the S3 bucket you’re downloading from

$ISISTESTDATA should now contain a full clone of the ISIS test data for running Makefile based tests.

Notes:

• Users can download specific files from the bucket by adding path data or file information to the first argument, that is, to download only the ‘base’ folder from the isis_testData bucket, the user could call: rclone sync remote:asc-isisdata/isis_testData/base
• It is important that users understand the difference in rclone’s ‘sync’ and ‘copy’ methods. ‘copy’ will overwrite all data in the destination with data from source. ‘sync’ replaces only changed data.
• Syncing / copying in either direction (local -> remote or remote -> local) results in any changed data being overwritten. There is no warning message on overwrite.

Installing older versions of ISIS

How do I install ISIS2?

If you are looking for ISIS2, please refer to the ISIS 2 Installation Guide for instructions on downloading and installing ISIS 2.

How do I install ISIS3.5.2 or earlier?

If you are looking for a version of ISIS prior to 3.6.0, please refer to the Legacy ISIS3 Installation Guide for instructions on downloading and installing ISIS, versions prior to 3.6.0

How do I access the ISISDATA download script with ISIS 7.0.0 or earlier

You can download the script and config file from the repo:

# install rclone 
conda install -c conda-forge rclone

# download the script and rclone config file
curl -LJO https://github.com/USGS-Astrogeology/ISIS3/raw/dev/isis/scripts/downloadIsisData

curl -LJO https://github.com/USGS-Astrogeology/ISIS3/raw/dev/isis/config/rclone.conf

# run the script as normal, using --config to point to where you downloaded the config file 
python downloadIsisData --config rclone.conf <mission> $ISISDATA

The script does not support python2, sometimes you need to explicitly use python3 with python3 downloadIsisData <mission> $ISISDATA --config rclone.conf

Semantic Versioning and Its Role in Describing the Software

In 2019, the ISIS project adopted semantic versioning via its second Request for Comment (RFC). Semantic versioning was adopted as a tool to help quickly describe how changes to the software impact users and developers. Versions of ISIS are now using a Major.Minor.Bug scheme (e.g., 7.1.0).

The Major, Minor and Bug numbers are in order of importance. The final (Bug) number is incremented whenever one or more bug fixes are included in a version. Neither users nor developers should see any changes in the way ISIS programs are called or how the API operates as the final (Bug) number increments.

The first two numbers indicate whether the change(s) are breaking or non-breaking. What is a breaking change? If a change to the API, defined as programs (e.g., spiceinit or cam2map or pds2isis) and some text output (e.g., CSV output, but not .txt), alters how a user calls the program or parses the program output in a way that a scripted solution would fail, that change would be considered a breaking change. In other words, if a CSV output file removed or renamed a column, that would be breaking. If a CSV file added a new column, that would not be breaking. Likewise, if an application spiceinit adds a new argument, that is non-breaking. If the change removes, reorders, or changes how the application (CLI) is called, the change is breaking.

Users

Users of ISIS benefit from semantic versioning because they can quickly determine whether or not an upgrade of their current version could include changes that would be breaking. When deciding whether or not to upgrade, users can safely assume that an upgrade of the minor version number will only add capabilities. Users should be more cautious with changes to the major version, as some breaking change(s) are included. How should a user proceed? Users should reference the Changelog to understand what changes have been made that necessitated an increase in the Major version number.

Developers

Developers writing against the ISIS API or writing code for submission to the ISIS project also benefit from semantic versioning. For the former use case, writing against ISIS, developer concerns are similar to user concerns. When has the API made of command line tools and program outputs changed? Does that change impact my pipeline or code? Do I need to adjust my work before updating versions (for example, to gain access to new features)? These questions are answered by checking the versioning and the Changelog.

Developers should ensure that changes that break the API are well-marked. Before making a breaking change to the API, we require an RFC to solicit input from the broader community. The RFC process allows impacted persons to discuss the change, propose alterations, and finally adopt or pause the inclusion of the change in the code base.

What update cadence does the project anticipate from users and developers?

The project is in the process of adopting a Long Term Support(LTS) model. Once fully adopted, the project assumes that either (1) users and developers will freeze the version they are using with no expectation of updates or (2) users and developers will update at either each LTS version increment (updating every 18 months) or work on the quarterly release (therefore updating every 3 months). Users and developers using the LTS or current release versions will benefit from bug fixes and new non-API breaking features.