UbiOps takes care of deploying and serving your data science code. It handles containerization of your code, deploying it on servers, request handling, automatic scaling, monitoring, managing data flows and outgoing connections to data storage.
More information about UbiOps
If you are looking for a high level description of the UbiOps platform and its purpose, take a look at our website.
UbiOps is built around the following core components:
Models are objects in UbiOps that serve a user's Python code. They can receive requests through the UbiOps API to transform input data into output data.
Connectors are configurable components to insert the output from your Models into a database or other data storage solution.
Pipelines are sequences of Models and Connectors. In a pipeline, a user can connect the output of a model to other models and connectors. Pipelines can receive requests like models. UbiOps will manage the data flow.
The best way to learn about UbiOps is to actually start using it. The Model deployment quickstart will teach you everything you need to know about how to deploy and serve your code on UbiOps. The Pipeline deployment quickstart covers more advanced functionalities in UbiOps.
Managing Projects, Users and Permissions¶
Every UbiOps user is part of an organization. An organization can have one or multiple projects. These projects contain resources like models, pipelines and connectors. See Organizations & Identity Management to learn more.
Within a project you can also create service users to obtain API Tokens to authenticate with the UbiOps API.
Roles can be assigned to both users and service users to manage permissions on various levels. For instance to grant or revoke access to projects and resources. To learn more about assigning permissions and roles, see Permissions and roles.
Interacting with UbiOps¶
There are multiple ways to interact with UbiOps. First of all, there's the WebApp which provides an easy way of managing and deploying your resources. However, if you need to programmatically interact with our platform, it is possible to do so via the API, Client libraries or Command Line Interface. All of these options offer the same functionality. You could, for example, create a new project using the Command Line Interface and edit it afterwards using the WebApp.