From the perspective of the web, each entity would have just ID and arbitrary context (dict), which would allow to ignore another details of the domain and simplify persistence (only two core-related DB models needed: Entity (id + content) and Action (id + order/parent + content).
This is probably not possible, because the web API needs to know more about the entities and because we need more structured DB entities for efficient search.
From the perspective of the web, each entity would have just ID and arbitrary context (dict), which would allow to ignore another details of the domain and simplify persistence (only two core-related DB models needed: Entity (id + content) and Action (id + order/parent + content).