Pipeline

A Pipeline is a set of pre-defined and re-usable actions called steps that can be executed manually or automatically. Pipeline might be composed of various steps such as queries, transformation, machine learning model training or inference logic, ...

Pipelines can be used to achieve many different purposes: e.g. execute a function to calculate a data point from others, training a machine learning solution, predict asset and/or process behavior, forecast and analyze some trends, trigger an alert, ...

A pipeline can be defined through the UI or using Python with the DS API. A pipeline is stored as a JSON file.

Steps

Pipelines are composed of different elements called steps. Each Step defines a sub-task within the pipeline, the step logic is re-usable accros pipelines and but can have its own configuration within a specific pipeline : e.g. you can define a linear regression training script using a coefficient that is specified for each pipeline using it.

There are different steps type such as : query, script, model, write, plot, ... Each steps are written in python, you can developed your own steps or used pre-defined one embedded within Wizata.

Templates

A pipeline can be based on a Template or not. When a pipeline uses a template, it can define a common behavior between multiple similar assets and/or processes : e.g. you can define a pipeline which trains a model to detect anomalies on motors and re-using it for each similar motors defined inside your Digital Twin.

Execution

Once a request to execute a pipeline is filled it produces an Execution. A Pipeline can be executed on different mode:

• Experiment : experiment mode is used to try out and research your pipeline. Experiment mode minimize the resource impact on your platform and uses different Runners
• Production : on production mode, Engine uses production Runners to execute your pipeline using optimal and dedicated resources.