Citrine v3.26.0 is released!

In this release of Citrine Python, we’ve added new capabilities to help you better manage your projects, including the ability to archive, restore, and view their archive status. We’ve also made several internal improvements and dependency updates to keep things running smoothly. This update also includes a fix for how archived assets are handled in DesignSpace listings. Finally, we introduce a deprecation for Predictor Evaluation Workflows in favor of a simpler method to execute evaluations directly without a registered workflow. Read more in our FAQ documentation.

What's New

• Add capability to archive/restore Projects and view archive status. #998, #1000

Improvements

• Improvements to our internal testing methods. #999
• Dependency updates. #996

Fixes

• Fix to DesignSpace listing endpoints to properly handle archived assets correctly. #1001

Deprecations

• Deprecate methods around Predictor Evaluation Workflows. Predictor Evaluations can now be triggered directly against a predictor, without the need to register a Predictor Evaluation Workflow. See our documentation for more details on how to update your code for this change. Predictor Evaluation Workflows will be removed in Citrine Python v4.0. #997, #1002, #1003

Full Changelog: v3.22.1...v3.26.0

Citrine v3.22.1 is released!

In this release of Citrine Platform, we're excited to provide new ways to filter data and AI resources. By providing new GEMD Query criteria, you can now filter Materials by Tag and by relationships with other processes. Additionally, by using the "Read-Only" status information for Design Spaces, you can now surface if your Design Spaces are Read-Only and why. We also now store Design Space configuration settings so they can be used by our web UI and future endpoints.

What's New

• Add new criteria enabling filter-by-tags using GEMD Queries: TagsCriteria and criteria to filter based on connectivity: ConnectivityClassCriteria. The TagsCriteria accepts 3 different TagFilterTypes for AND, OR, and NOT logic. #991
• Expose locked metadata for Read-Only Search Spaces. #993

Improvements

• Design Spaces constructed with the create_default method will store configuration settings and the Predictor used for configuration, a better web-app experience. #992
• Updated dependencies. #994

Full Changelog: v3.19.0...v3.22.1

Citrine v3.19.0 is released!

In this release of Citrine Python, we are excited to introduce Default Labels! These labels can be applied to Materials (as opposed to only Ingredient Specs) and are inherited by default for any new Ingredients referencing that Material. This makes it easier for you to label one object rather than a group of ingredients as well as enforce labeling of given ingredients.

What's New

• We now support Default Labels on Material Objects. By using the default_labels argument on Material Specs, you can apply a list of strings to these objects that will act as defaults for future Ingredients. Any future Ingredient Specs registered from that Material Spec will inherit that set of labels unless labels are otherwise specified. default_labels can also be applied to Material Runs and will be leveraged in any design workflows for which that Material Run is a component. This is very handy when you have an material that does not exist in any of your final formulations, but can be tested with a predictor. #989, #990

Full Changelog: v3.17.0...v3.19.0

Citrine v3.17.0 is released!

In this release of Citrine Python, we've introduced capabilities to further streamline our users' AI-driven explorations. Users can now directly access Feature Effect values used in to create the Feature Impact plots in our web UI for each model output, enriching their understanding of the model's predictions. Additionally, previously UI-exclusive Candidate attributes such as pinned and hidden attributes, candidate name, and comments, are now viewable in a read-only format within Citrine Python, offering a more options for custom automated filtering and visualization. We've also included some updates to enhance our own development and experimentation with platform assets, ensuring we continue to deliver best-in-class materials AI capabilities.

What's New

• Support for accessing Feature Effect values from valid Graph Predictors. The call predictor.feature_effects() will return a FeatureEffects object with feature effect values for each output of the chosen predictor. #982
• Expose Candidate attributes previously only seen in the UI, including pinned and hidden attributes, candidate name, and any candidate Comments. Attributes are read-only in the context of Citrine python. #987, #988

Improvements

• Expose the experimental AttributeAccumulationPredictor node for internal development and testing. #984

Fixes

• Improve unit test efficiency. #986

Full Changelog: v3.11.6...v3.17.0

Citrine v3.11.6 is released!

We are pleased to roll out the latest version of Citrine Python, featuring significant documentation improvements and an important bug fix that enhances overall functionality for managing projects. We will continue to keep our documentation up to date and squashing bugs keep you all running smoothly!

Improvements

• Improvements to our documentation. #972

Fixes

• Bug fix to find_or_create and similar methods to respect team locations. #974, #975, #976, #977, #978, #980

Full Changelog: v3.11.0...v3.11.6

Citrine v3.11.0 is released!

We're excited to announce the latest enhancements to Citrine Python that continue our commitment to providing our users with a robust and efficient experience. In this release, we've focused on improving our table configuration processes and optimizing our ingestion workflows.

As our platform evolves, we are incorporating faster and easier ways for users to define their training sets, and we now connect our power users to those methods. Users can now leverage direct access to our advanced default table definition processes that leverage GemdQuery objects for even smoother table builds.

We've also taken steps to ensure that your data ingestion is not just effective, easy to use, with a direct link from data ingestion to training-set builds inside of a project. For those who value up-to-date and clear documentation, we've made corrections and enhancements to ensure that your integration with our SDK is as smooth as possible. Additionally, this release includes several internal improvements aimed at streamlining development, allowing us to consistently deliver the reliability and performance you expect from Citrine.

What's New

• Provide direct access to our latest default table definition methods via from_query for constructing a Table Config. #967

Improvements

• Modify compound ingest operation to wait for table build completion #970
• Internal improvements to streamline development #963, #964, #965, #968

Fixes

• Corrections to our documentation #966
• Add project argument to ingestion routes to enable table builds in the same call as ingestion #969

Full Changelog: v3.5.3...v3.11.0

Citrine v3.5.3 is released!

In this release of Citrine Python, we have a few bug fixes to improve our user experience. We've also deprecated the training_data field on our sub-predictors, as those have been required to match the root graph predictor for quite some time. And we've been keeping our dependencies in line, to make sure you all can install and maintain your environment smoothly.

Fixes

• Update to dependencies, eliminating warnings. #961
• Fix reading file links in Windows environments. #950
• Correction to GEM Table creation warning strings. #962

Deprecated

• The training_data field on sub-predictors has been deprecated, as having a data source distinct from the root graph predictor would fail to register. #960

Full Changelog: v3.4.8...v3.5.3

Citrine v3.4.8 is released!

In this release of Citrine Python, we are excited to now support more ratio-type units, such as % and ppm, in our GEMD ingestion. And, per usual, continuous updates to address instabilities and keep you running smoothly!

What's New

• We've updated our gemd-python dependency to now support % and other ratio expressions as units. %, ppm, and other ratios are available here and on our platform. #959

Fixes

• Updated enumerated design spaces to restrict the allowed data value types. #958

Full Changelog: v3.4.6...v3.4.8

Citrine v3.4.6 is released!

In this version of Citrine Python, we are excited to introduce a pivotal change to the structure of assets on our platform. We know our users want to get every piece of value out of their data and one way to limit the value of your data that is to keep it locked up in silos. As part of the introduction of Data Manager in the Citrine Platform, we have taken Datasets out of Projects and made them assets of a Team, allowing users direct access to all of the data in their Team to leverage in AI Projects.

This introduces non-breaking changes, but our users are encouraged to migrate to new Team-based or Dataset-based endpoints for data management as soon as possible to (A) take advantage of the Data Manager feature on the Citrine Platform and (B) prepare for the eventual removal of Project-based endpoints. For more details on how this will affect your code, see the Migrating to Use Data Manager guide in the FAQ section of our documentation or reach our to your Citrine support team.

But that's not all we're bringing in this release. We've also updated our Molecular Generation package to leverage SMARTS notation in defining constraints and introduced simpler filtering methods in our listing methods of AI Assets. And as always, we are keeping our code up to date to maintain data security and keep our users running smoothly!

What's New

• New endpoints and deprecation warnings are introduced to support the use of the Data Manager feature. The key change is that newly registered Datasets are now assets of Teams rather than Projects. We have included new collection methods at the Team and Dataset level to account for this, while also adding deprecation warnings for Project-level collections that will be no longer supported in v4.0. For more details on how this will affect your code, see the Migrating to Use Data Manager guide in the FAQ section of our documentation or reach our to your Citrine support team. #947, #949, #951, #952, #953, #956

Improvements

• Updates to our Generative Molecular Design package to use SMARTS format for con straining the substructure in a generative design execution #939
• A simpler filtering strategy for listing Predictors and Design Spaces #947

Fixes

• Fix to our PyPi Deployment #957
• Internal maintenance and fixes #945, #946, #948, #953

Full Changelog: v3.2.11...v3.4.6

Citrine v3.4.4 is released!

In this version of Citrine Python, we are excited to introduce a pivotal change to the structure of assets on our platform. We know our users want to get every piece of value out of their data and one way to limit the value of your data that is to keep it locked up in silos. As part of the introduction of Data Manager in the Citrine Platform, we have taken Datasets out of Projects and made them assets of a Team, allowing users direct access to all of the data in their Team to leverage in AI Projects.

This introduces non-breaking changes, but our users are encouraged to migrate to new Team-based or Dataset-based endpoints for data management as soon as possible to (A) take advantage of the Data Manager feature on the Citrine Platform and (B) prepare for the eventual removal of Project-based endpoints. For more details on how this will affect your code, see the Migrating to Use Data Manager guide in the FAQ section of our documentation or reach our to your Citrine support team.

But that's not all we're bringing in this release. We've also updated our Molecular Generation package to leverage SMARTS notation in defining constraints and introduced simpler filtering methods in our listing methods of AI Assets. And as always, we are keeping our code up to date to maintain data security and keep our users running smoothly!

What's New

• New endpoints and deprecation warnings are introduced to support the use of the Data Manager feature. The key change is that newly registered Datasets are now assets of Teams rather than Projects. We have included new collection methods at the Team and Dataset level to account for this, while also adding deprecation warnings for Project-level collections that will be no longer supported in v4.0. For more details on how this will affect your code, see the Migrating to Use Data Manager guide in the FAQ section of our documentation or reach our to your Citrine support team. #947, #949, #951, #952, #953

Improvements

• Updates to our Generative Molecular Design package to use SMARTS format for constraining the substructure in a generative design execution #939
• A simpler filtering strategy for listing Predictors and Design Spaces #947

Fixes

• Internal maintenance and fixes #945, #946, #948, #953

Full Changelog: v3.2.11...v3.4.4