django-river-ml

What's Changed

• Update river by @vsoch in https://github.com/vsoch/django-river-ml/pull/17

Full Changelog: https://github.com/vsoch/django-river-ml/compare/0.0.2...0.0.21

What's Changed

• Add minimal support for creme models by @vsoch in https://github.com/vsoch/django-river-ml/pull/15

Full Changelog: https://github.com/vsoch/django-river-ml/compare/0.0.19...0.0.2

What's Changed

• Adding functions to save and load local pickle files by @vsoch in https://github.com/vsoch/django-river-ml/pull/14

Full Changelog: https://github.com/vsoch/django-river-ml/compare/0.0.18...0.0.19

This release adds a basic visualization to the example app to look at cluster centroids, and an iris dataset example (using scikit learn datasets).

What's Changed

• adding example view and example with iris dataset by @vsoch in https://github.com/vsoch/django-river-ml/pull/13

Full Changelog: https://github.com/vsoch/django-river-ml/compare/0.0.17...0.0.18

• river.metrics.cluster was removed, so we cannot use them here.

This release includes:

• refactor of client to provide same functions internally
• allow user to specify a custom set of urls to expose
• documentation for DjangoClient (internal client)

This release includes a bug fix for download (the URLs were out of order so a download request was hitting the wrong endpoint) and support to disable rate limiting globally.

This release is a bugfix to ensure the /auth/token url is in the same namespace given to django river ML. E.g., /auth/token is commonly used by django applications and in testing I just hit a bug where my auth was failing because django River ML stole the endpoint! 😆

This release includes support for custom and cluster models.

This is the first automated release of Django River ML! 🥳

Specifically, this is the first release with examples for all model types (binary, regression, multiclass) and the addition of a label endpoint to be able to save an identifier for a prediction and then label it later. See https://github.com/vsoch/django-river-ml/pull/7.