Overview

Jubakit is a Python module to access Jubatus features easily. The goal of Jubakit is to:

• Accelerate cycles of data analysis by integrating powerful scikit-learn features into Jubatus
• Conceal the system architecture of Jubatus (config files, TCP port number assignment, process invocation, RPC call, etc.) to provide an interface that users can focus on data analysis

Here is a shortest example of code using Jubakit. You can perform anomaly detection on CSV dataset only by 4 lines of Python code:

dataset = Dataset(CSVLoader("dataset.csv"))
service = Anomaly.run(Config())
for result in service.add(dataset):
  print(result)

Jubakit provides a simple-to-use APIs while allowing users to customize detailed behaviors. Jubakit also comes with configuration parameters that works well in most cases, so you don’t have to configure them until necessary.