Data & Objects

Once preprocessed, your data is stored in the data storage of the pipeline. This data is now in the form of objects. Each object represents a single unit of data that can be used for text generation.

Examples of objects could be:

• products
• product variants
• categories
• places

The only mandatory attribute of an object is uid, which uniquely identifies the object in the data storage. The uid only needs to be unique inside one pipeline.

In the data section of the pipeline you can view all stored objects in a table and filter and analyze them with our powerful query UI.

If you have autogenerate enabled for your data pool, any newly created or modified objects will automatically start a new generation job. If a preprocessed upload does not actually change the data of an existing object, no new generation job will be started.

Studio vs Cockpit

In contrast to documents in Cockpit, objects in Studio only save your structured data, and do not include any generation results. These are saved separately in the results section of the pipeline.

More to Come

We currently only offer the contains operator as a filter. Don't worry though, we will add more filter options soon.

We also plan on introducing data pools for different types of data. This concept will be completely different to collections in Cockpit, and will not be used to control generation.

TODO multilingual