Here's what's new in Release 1.1.2 (8 November 2017):
Bug fixes
Fixed an issue with duplicate requests when running the StockViewer sample PixieApp as a web application ( #483 ).
PixieDust can now run the same code in multiple notebooks locally with Spark. Previously, running the same code across multiple notebooks locally, such as PixieDust's
sampleData()
API, generated an error (
#496
).
Refactored PixieGateway into its own GitHub repo to run in IBM's Data Science Experience environment and updated PixieGateway documentation ( #498 and #495 ).
Enhancements
Released alpha version of PixieGateway admin dashboard, with a streamlined UI. DevOps users of PixieGateway can now access a list of PixieApps deployed to the gateway and a list of running kernels with real-time statistics ( #487 and #501 ).
Added the ability to
display()
JSON directly. PixieDust now supports native rendering of data dictionaries, or an array of dictionaries, without having to first convert them into Pandas DataFrames (
#489
).
Added functionality to embed PixieDust charts into a regular web application, such as a web app that was not generated via PixieApp web publishing ( #493 ).
Improved PixieApp web publishing to allow developers to publish to a specific kernel. Previously, all PixieApps were published into the default Python kernel ( #486 ).