Generation Workflow

We have introduced workflows to guide users through various use cases. The Generation Workflow offers step-by-step directions through the process of generating multi-model data from a single model input. It encompasses the complete process, including the inference of a schema from raw data, editing the inferred schema category, creating mappings, and generating the multi-model data itself. Following steps will guide you through the process:

The Home page of the MM-cat client presents the user with multiple options. To start the workflow process, fill in the workflow label and press “Confirm”.

Select or create a new Datasource. When creating new, fill in the required information. Press “Continue”.

The Inference Job has run. If succesful, the state badge will change to “Waiting”. Edit the inferred Schema Category to your liking.

Define the edit and see a live preview of it by pressing “Confirm”. To keep the change press “Save”, otherwise press “Cancel”. All of the saved edits stay available in the Inference Editor, you can choose to undo or redo them any time. Once you are done with editing, press “Save and Finish” followed by “Continue”. See below for full description of Inference Editor.

The Inference Editor provides multiple editing options, affecting both the structure (Merge, Cluster, Recursion) and the layout (Layout, Save Positions) of the inferred Schema Category. Click through the context menu to discover more.

Choose from the available layouts the one that best fits your schema. Change your choice even while schema structure editing.

This structure editing operation enables you to identify objects which define relationships in the schema. Choose to define a Reference or a Primary Key. Both of these options provide a Manual or a Candidate choice. Click on objects to define the edit manually or select from a list of candidates. See an example of suggested candidates below, see manual choice in the picture in Step 2.

The Cluster structure editing operation simplifies the schema by creating a cluster representant for multiple cluster members. A schema contains a cluster if there are multiple objects which have the same structure and name, or their names share a common cluster identifier. Identify these objects by clicking on them.

Recursive schemas can be simplified by the Recursion structure editing operation. Define the recursive pattern in your schema by clicking objects and morphisms.

View the saved edits and undo or redo them as you like.

If you wish to manually adjust the schema layout, do so by dragging the graph nodes around. Once you are done, save the new positions by clicking this button.

Define at least one output Mapping. First, select the output Datasource. Then define the Mappings. You can either load an initial Mapping or create a new one from scratch. Both options enable you to edit the Mapping as you go in the Mapping Editor using the context menu or the keyboard shortcuts. Press “Finish Mapping” when your done defining your output Mappings. To see the results press “Continue”.

Depending on the type of the output Datasource, you will see either Data Manipulation Language (DML) commands or generated files.

Enjoy your generated multi-model dataset!