Good Tables ships with a set of core processor classes.
All processors can be used as standalone processors, as well as being part of a pipeline.
Validates the data structure conforms to expectations for rows and columns.
Validates the data against a JSON Table Schema.
If no schema is passed, optionally creates a schema from the data.
When working with Good Tables as a library, it is possible to call custom processors in your pipeline.
We have the start of some documentation of this in the Tutorials section.
A better entry point now may be to see the existing processors and use those as a starting point for implementing your own.