Citrine v1.41.0 is released!

In this release of Citrine Python, we've added some additional functionality that will enable users to share Predictor Evaluation Workflows (and their results) within a Team. We've also removed official support for python 3.6, which reached end of life in late 2021, to keep up with the steady drum beat of software development.

What's New

• Users can now publish Predictor Evaluation Workflows (PEWs) from a project to a team and pull PEWs into other projects within that team. Doing so allows users to view PEW results from any project with access to it and the predictor that was evaluated. #784

Deprecation

• Remove official support for python 3.6. #783

Full Changelog: v1.40.0...v1.41.0

Citrine v1.40.0 is released!

In this release of Citrine Python, we are shipping a new tool while paving the way for future functionality. We are proud to introduce the Holdout Set Evaluator, which allows users to evaluate model performance on a user-defined holdout set in lieu of our typical cross validation strategy. We've also included updates to our python SDK that will eventually allow users to select different algorithms in their AutoML Predictors and handle archival of predictors once they have been properly versioned. Both features are still in development for most production deployments, but these changes allow us to test and iterate before full deployment.

What's New

• Preparation for algorithmic selection in AutoML Predictors. The addition of the estimators field will eventually allow users to select additional algorithms to be considered during training. Full backend functionality still to come. #780
• Introduction of the Holdout Set Evaluator for generating model error metrics on a customizable holdout set in lieu of typical cross validation. The HoldoutSetEvaluator, which can be used in a Predictor Evaluation Workflow alongside or instead of a CrossValidationEvaluator, takes a Data Source as an argument that is then predicted with the model during workflow execution. The same set of model performance metrics, such as RMSE and PvA results, can then be calculated and returned in the execution results. #768

Improvements

• Updated reported predictor archival status to prepare for a change with predictor versions. #782

Full Changelog: v1.37.1...v1.40.0

Citrine v1.37.1 is released!

In this release of Citrine Python, we are shipping internal improvements to both enable further development and allow for smoother debugging in the future. As always, we are constantly improving Citrine Python to be more reliable and flexible to meet your needs.

Improvements

• Additional fields added to predictor reports to support future feature delivery. #778

Fixes

• Introduce BATCH_SIZE as a package constant that to control how many objects are registered to the platform at one time. This constant can be extended to debug issues in data service timeouts. #779

Full Changelog: v1.36.6...v1.37.1

Citrine v1.36.6 is released!

In this release of Citrine Python, we are providing minor fixes and improvements to keep our users running smoothly and building confidently. We've improved our documentation for contributions and included fixes to some of our deprecation warnings brought about by the Teams migration.

Improvements

• Improved documentation for contributions. #775

Fixes

• Correct deprecation warning on find_or_create_project and projects.register to trigger only when used incorrectly in the Teams context. #774, #777
• Internal fix to default design space creation. #776

Full Changelog: v1.36.2...v1.36.6

Citrine v1.36.2 is released!

In this release of Citrine Python, we are adding more functionality for the upcoming migration to Teams, including the ability to quickly access user permissions and resources available to a project at the Team level. We've also improved our search functionality around files and performed some additional under-the-hood work to fix bugs and improve performance.

What's New

• New endpoints to allow quick access of a user's permissions on a given Team. Using team.me() will allow you to check your permissions within a team with one call! #770
• Added functionality to list the IDs of accessible resources at a Team level. This allows you to see what assets are available to use in their project that have been published from other projects in your team. #769

Improvements

• Improved the way we search for files on our platform by using filters. #772

Fixes

• Minor improvements and bug fixes. #771, #773

Full Changelog: v1.34.0...v1.36.2

Citrine v1.34.0 is released!

In this release of Citrine Python, we are happy to announce additional configurability of Design Spaces that are created via the create_default method. The method, which creates a new Design Space based on a given Predictor configuration, now takes three optional flags: one each to constrain the design space ingredient amounts, label amounts, or ingredient-counts-per-label based on the shape of the training data.

This lets you build your default design space in a number of ways: you can start similar to your training data and expand, or start with a big design space and constrain smaller. All these flags are optional; omitting them will default to "False" resulting in an unconstrained design space. As always, we are also shipping minor bug fixes to keep you running smoothly.

What's New

• Add configurability flags to the create_default Design Space method. #767

Fixes

• Minor bug fixes and improvements. #766

Full Changelog: v1.33.2...v1.34.0

Citrine v1.33.2 is released!

This release is a quick fix to some formatting issues with our deprecation warnings around project and user management methods affected by the migration to Teams for access control. The fixes should increase the readability of the warnings and make it easier to know where and how to prepare your code for migration. There is no change to what methods will be deprecated or replaced in the migration.

Fixes

• Formatting fix to deprecation warnings around project and user management. #765

Full Changelog: v1.33.1...v1.33.2

Citrine v1.33.1 is released!

In this release of Citrine Python, we've made one small, but important addition, to our warning system. Any methods around project and user management that will be deprecated with the migration to Teams for access control are now accompanied by a readable deprecation warning pointing you to the post-migration equivalent. This will allow you to test your code before migration and point you to the new methods to support your workflow once migration has been completed.

What's New

• Added deprecation warnings for methods to be deprecated in the migration to Teams for access control. #763

Full Changelog: v1.33.0...v1.33.1

Citrine v1.33.0 is released!

In this release of Citrine Python, we are all about improvements. We've made improvements to our default behavior around units in GEM Table column definitions to avoid the case of miss-matched units within a column. We also have improved backwards compatibility of project endpoints to reduce the amount of code changes required in the upcoming migration to Teams. Finally, as always, we continue to update our documentation and make minor improvements to better support our users.

Improvements

• Smarter behavior of the target_units parameter in GEM Table Column definition is now on platform. If you don't specify which units you want to work in, the table builder check if there's also an OriginalUnitsColumnDefinition to figure out if you wanted the numbers as reported; if not, you'll get the default_units from your Attribute Template. So if you don't specify units, the platform's got you covered. #758
• Increased backwards compatibility of project endpoints for our upcoming migration to Teams for access control. #759, #761
• Updated our documentation to be more descriptive about the relationship between branches, design workflows, and archival. #760
• Other minor improvements. #762

Full Changelog: v1.31.1...v1.33.0

Citrine v1.31.1 is released!

This release includes exiting new functionality for building predictors. For fans of the SimpleMLPredictor, we've included the simple_ml_builder that mimics the simple interface, but creates a GraphPredictor that is compatible with our web UI. We also now allow more flexibility in referencing files through our SDK. As always, we are improving our documentation, performance, and fixing bugs to keep our users running smoothly.

What's New

• We've introduced the simple_ml_builder: the ease of use of the SimpleMLPredictor but directly compatible with the current version of our Model Builder web UI. One can use this builder method to pass input, output, and latent variable descriptors and get a GraphPredictor with the necessary AutoML and featurizer nodes needed to predict your outputs. #737, #752, #755
• Users can now retrieve files by the file_link object, its UUID, or its filename. #745

Improvements

• Improving our documentation for querying and waiting for asynchronous tasks. #749
• Updated dependency to latest release of gemd-python (see gemd-python release notes for details). #757
• Performance improvements to accelerate our internal development. #751, #756

Fixes

• Bugfixes to keep folks running smoothly. #750

Full Changelog: v1.28.3...v1.31.1