Release Notes

Release notes will mention the release date, a summary for each release, and comprehensive descriptions of major changes, minor changes, feature removal, and bug-fixes. This is typically more comprehensive than the announcements on the GitHub Discussion board: https://github.com/orgs/MAAP-Project/discussions/categories/announcements.

4.2.0

March 3, 2025
Release with several important improvements. To use the new features, please start a new v4.2.0 workspace.

Breaking Changes

  • Updated most packages in all four base images. The Pangeo, isce3, and python images are pinned to versions in pangeo-notebook 2024.11.11 version. Important package updates are:
  • All images running on MAAP will now be using mini-forge to avoid licensing issues. Impacts:

    • All conda installs will now pull from conda-forge

    • No longer using R channel on anaconda (also continuing to not use the defaults channel)

    • If you cannot find your package on conda-forge or another open source community channel, we recommend installing your package from their install instructions on their GitHub README or contact the platform team.

    • Anaconda’s new terms: https://www.anaconda.com/pricing/terms-of-service-faqs

Added

  • Container URL in Algorithm Registration UI is now a dropdown where the default is the maap_base image which is the fastest for algorithm registration

    • Users who would like to use this faster maap_base option must self manage their packages (i.e. in an environment.yml). More information in the docs here

  • Added more packages requested by users for the R image, especially Geotrees specific packages like lasR, partykit, BIOMASS and tidyterra

  • The Algorithm-Registration UI input-boxes are narrower to help reduce the need for horizontal scrolling

  • Improvements to the Jobs UI: added job duration and more a obvious cancel button, and cleaned up the filter controls

  • Using the maap_base image as the base image for all workspace images

Fixed

  • Public ssh key is now correctly being pulled from MAAP profile page into /projects/.ssh/authorized_keys upon workspace launch if it was not already present

  • Display SSH info giving undefined error resolved

  • Fixed status filtering issues for Jobs UI

  • Fixed a broken documentation link that shows up during algorithm registration

4.1.1

October 23, 2024
Minor release with non-breaking changes. To use the new features, please start a new v4.1.1 workspace.

Fixed

  • Removed pygeos from all base images.

  • Increased number of records in the Jupyter Jobs UI to 200.

  • Expanded s3 permissions for the maap-py aws.workspace_bucket_credentials function. Credentials for user bucket folders now include full read/write access. See related documentation.

4.1.0

October 2, 2024
Release with several important improvements. To use the new features, please start a new v4.1.0 workspace.

Significant Changes

  • A new maap_base image container is available with a minimal runtime environment for use within DPS. Read more below.

  • A way to securely manage credentials for other services (e.g., Google Earth Engine credentials) and pass them into DPS Jobs. This is called Secrets management.

  • Organizations feature to help enable dedicated compute resources for user groups and prevent unauthorized access to resource queues.

  • MAAP DPS Sandbox queue for quick testing and validating algorithm builds in DPS.

  • Some older algorithms may require re-registration due to underlying DPS software being upgraded.

Breaking Changes

  • Please migrate to the newer version of workspaces as soon as possible in order to leverage the improved security.

  • The platform team will help migrate Algorithms to the optimized base images, over the next three months.

  • Some older Algorithms may see an error stating /home/ops/.maap-dps.env file not found or /app/dps_wrapper.sh: line 10: python: command not found when running on the DPS. This requires re-registration of the Algorithm.

Added

  • Organizations features

    • Added new Organizations features and user access-control to DPS job-queues. You will now only see the Job queues that are available to you; if you are missing a queue please ask a team lead or the platform team for help configuring your permissions.

  • Secrets Management

    • Added ability to store Secrets using maap-py and utilize them within DPS Jobs. This will help accommodate things like passing Google Earth Engine credentials to your Algorithm in a secure way.

  • New maap_base image

    • New maap_base image that will speed up Algorithm Registration and smaller DPS Jobs significantly containing just conda by reducing the time for the system to build the runtime environment. Smaller jobs will also run more quickly.

    • In order to use this new container, algorithm build scripts must specify all of the required libraries in their conda environment.yaml file.

    • To use this new image, specify the following url during algorithm registration mas.maap-project.org/root/maap-workspaces/custom_images/maap_base:v4.1.0.

  • MAAP DPS Sandbox queue

    • A new queue maap-dps-sandbox has been added for quick testing of registered algorithms.

    • This new queue has limited resources (8gb), and a max run time of 10 minutes for algorithms.

    • The motivation for the queue is for users to quickly test their algorithm build, conda env, run script, input handling, etc.

    • This is the only queue available to Guest accounts

  • DPS software upgraded to HySDS v5.2.0

Fixed

  • Upon registering an Algorithm using the Registration UI, the build-link now opens in a new tab. It also provides the path to where the config yaml file is stored in your workspace and a notification with the registration link and algorithm name:branch.

  • Better error-handling when an Algorithm fails to register.

Security Improvements

  • Enhanced security for the DPS job-management workspace to check if a user is logged in.

Known Issues

  • Some older algorithms failing

    • Some older algorithms may see an error stating /home/ops/.maap-dps.env file not found or /app/dps_wrapper.sh: line 10: python: command not found when running on the DPS.

    • This is known to happen when the same github repository is used for registering multiple algorithms using the same branch (eg. main or master). This can be resolved by re-registering your algorithm.

    • For the future, if you want an algorithm container to stay unchanged, consider using github tags and registering an algorithm from that tag (eg v1, v2, etc).

FAQs

  • I dont see the resource queue I used in the past

    • Check if you can use any alternative available queues.

    • If not, contact the platform team to request your specific queue.

  • My algorithm is now failing

    • If your previously successful algorithm is now failing, try re-registering once again.

    • If for any reason re-registering is not an option reach out to the Platform Team.

4.0.0

July 3, 2024
Major new release with breaking changes.

Breaking Changes

  • The “Basic Stable” workspace stack has been renamed to “Python (default)”. The associated default vanilla conda environment has been renamed to python.

Workspace impacts:

  • This does not impact the r or icse3 workspaces.

  • If you are using the Basic Stable (vanilla) workspace, please upgrade to the new Python (default) workspace.

  • To use the new workspace to run a vanilla algorithm (e.g., running your existing algorithm in the Jupyter Terminal), you will need to update your scripts to reference the python conda environment instead of vanilla.

DPS Algorithm Impacts:

  • To use the new Python (default) as the base image for your DPS runs, you will need to update your build scripts and environment.yml files to use python instead of vanilla. Additionally, your run script will need to make sure it’s running in the python conda environment (not vanilla).

  • DPS algorithms being registered will use the new container URL mas.maap-project.org/root/maap-workspaces/base_images/python:v4.0.0 — also ensure that they are using python as the conda-environment name and not vanilla.

Added

  • JupyterLab has been updated to v4.

  • All extensions are available to be installed via pip now.

  • Added lonboard to all images.

  • Added all packages from the previous vanilla workspace (now called python) to the R workspace.

  • Maap-py no longer tracks a maap.cfg file. When using maap.py you should no longer indicate the API endpoint URL (maap = MAAP() instead of maap = MAAP(api.maap-project.org)).

  • Integrated Playwright testing framework into custom JupyterLab extensions.

  • Added support for the “maap-data-reader” assumed role. This allows DPS jobs and workspaces direct s3 access to certain DAAC buckets without requiring credentials or manual token refreshing (see https://docs.maap-project.org/en/latest/technical_tutorials/access/direct_access.html).

  • Migrated MAAP API PostgreSQL instances to RDS.

Fixed

  • The default base image in the ADE algorithm registration form now shows the correct MAS environment and workspace type.

  • The API error “Client is not EDL policy compatible” is now included in the response from the maap-py method aws.earthdata_s3_credentials(url) when applicable.

3.1.5

April 1, 2024
Minor feature additions and bug fixes.

Added

Fixed

Changed

3.1.4

January 22, 2024
Minor feature additions and bug fixes.

Added

Fixed

Changed

3.1.3

November 2, 2023
Minor feature additions and bug fixes.

Added

Fixed

Changed

Removed

3.1.1

October 4, 2023
Minor feature additions and bug fixes.

Added

Fixed

Changed

Removed

  • isce2 workspaces are now deprecated.

3.1.0

August 4, 2023
Hotfix to handle bugs that make working in the “new” ops ADE difficult

Added

  • Added more capacity to the new ADE cluster to support more concurrent users.

Fixed

Changed

Removed