Release compatible with Tune Insight v1.6.0. Install with

pip install tuneinsight==1.6.0

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.6.0

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This releases introduces minor changes to the SDK:

• API updates for compatibility with Tune Insight v1.6.0.
• Effective deprecation of Maas and encrypted regression.
• Bug fix for _disable_patch that could cause a Project object to never update its state if get_computation is called when the project has no computation definition.
• Bug fixes and improvements to the TIQL query generation interface.

Release compatible with Tune Insight v1.5.0. Install with

pip install tuneinsight==1.5.0

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.5.0

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This release introduces major changes to the SDK:

• The smallest supported version of Python is now 3.10. Please upgrade if you are still using Python 3.9. Typing annotations have been updated accordingly.
• Support for fetching Computations from the API as client.Computation objects. This can be used to start computations asynchronously (Computation.run_async) and retrieve results later.
• Deprecation of PIR (private information retrieval), MaaS (Model as a service), and GWAS (Genome-wide association studies) as computations.
• New (SDK exclusive) Heatmap computation that computes a 2-way histogram.
• Improved support for TIQL (cross-standard queries): all queries in the language can now be written using syntax sugar in tiql.py.

Other smaller changes:

• Support has been added for datasource commands

Release compatible with Tune Insight v1.4.1 (there is no SDK v1.4.1). Install with

pip install tuneinsight==1.4.2

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.4.2

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This is a minor release intended for compatibility with fixes in v1.4.1., and is also compatible with v1.4.0 (although a warning will be raised about mismatching versions). It introduces the utils.debug module and its dump(obj) function that returns a sanitized JSON representation of an object to be sent for debugging purposes.

Release compatible with Tune Insight v1.4.0. Install with

pip install tuneinsight==1.4.0

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.4.0

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This PR introduces support for new and pre-existing features:

• Secure inference to run inference on encrypted data.
• Retrieving all columns available on a project from a Project (including parsing of the datasource query).
• New preprocessing operations (filling missing values and removing duplicates)
• Improved error messaging and minor bug fixes.

Minor patch release compatible with Tune Insight v1.3.1 and v1.3.0. Install with

pip install tuneinsight==1.3.1

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.3.1

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This is a minor release intended for compatibility with fixes in v1.3.1., and is fully compatible with v1.3.0.

Release compatible with Tune Insight v1.3.0. Install with

pip install tuneinsight==1.3.0

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.3.0

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This release includes quality-of-life improvements and bug fixes, aiming to improve the usability of the SDK:

• We have improved error messages and warnings for some common issues (e.g. errors with .login).
• The cryptolib has been trimmed down and is now much lighter, reducing the size of the package.

In addition, we have decided to deprecate model-as-a-service and fully encrypted regression in the SDK. Use HybridFL instead.

Release compatible with Tune Insight v1.2.3. Install with

pip install tuneinsight==1.2.3

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.2.3

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This release includes fairly significant updates to the API for compatibility with new features. Notable changes:

1. Screening

This release includes API updates for compatibility with new features (including data preparation and cross-standard queries). Changes to the SDK itself are very minor.

Release compatible with Tune Insight v1.2.0. Install with

pip install tuneinsight==1.2.0

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.2.0

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This release includes API updates for compatibility with new features. Notable changes include:

1. Privacy risk analysis on projects (Project.display_privacy_warnings).
2. Many bug fixes, pertaining to loading computations from the backend and compound preprocessing.
3. Small improvements to the cryptolib that make it lighter.

Release compatible with Tune Insight v1.1.4. Install with

pip install tuneinsight==1.1.4

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.1.4

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This release mostly includes API updates for compatibility with new features added in the solution. Two notable SDK-only changes:

1. Basic support for Cross-standard queries (TIQL).
2. Improved plotting of Kaplan-Meier survival curves.

Release compatible with Tune Insight v1.1.2. Install with

pip install tuneinsight==1.1.3

To also install data science tools such as jupyter, install the full release:

pip install tuneinsight[full]==1.1.3

(If you are using zsh, you may need to escape the brackets as tuneinsight\[full\]).

This release brings very few changes since many features were brought in the intermediate release v1.1.2. The most notable change is a fix to a bug where the cryptolib could not be loaded when running a Linux image with Docker on Mac OS.