I would like to have some feedback either positive or negative about this library. Some ideas about future improvements are:

- generating more specific fields by using `faker`,

- integration with orm layers (`mongoengine` for example) to guess the structure of the data instead of defining it.