9. Use pydantic for JSON (de-)serialization

Date: 2020-06-09

Status

Accepted

Context

In order to interact with a remote workspace, for example, at structurizr.com. The remote or local workspace has to be (de-)serialized from or to JSON.

Decision

In order to perform these operations we choose pydantic which has a nice API, active community, good data validation, helpful documentation, and good performance.

Consequences

We separate the models representing Structurizr entities and their business logic from how those models are (de-)serialized. That means that for each model we have a corresponding IO pydantic model describing the JSON data model.