Enabling tensorflow in data scientist workflow

highlights

• Now is possible use tensorflow in steps like create a model, train this model, and more!
• The function/python enables the operations doesn't supported in learningOrchestra, is possible use this feature between workflow steps.
• dataset/generic enable the storage of any dataset format available from an url, use the function/python to treat the dataset format to a proper format like numpy or dataframe.

Introducing data scientist pipeline

highlights

• Introducing the data scientist pipeline. Now the user can run all steps from a workflow using the learningOrchestra steps, there are steps that enable a python code execution, and also there are steps to run methods from popular libs as scikitLearn.
• With the data scientist pipeline feature, we remove the PCA and TSNE features, the user can run this methods using the /explore feature, enabling the execution from several methods from libs as sckitLearn, and automatically plotting the result.
• There are fixes to improve the learningOrchestra performance and to kill bugs.

The fisrt stable version of learningOrchestra

This release has the initial version of learningOrchesta.

highlights

• Several microservices to diferent steps of data scientist processing.
• Scalable spark workers to Machine Learning processing.
• Easy deploy in cluster environments, including the Cloud.
• Python client to facilitate the microservices utilization.
• Gateway API to simplify the REST API usage.