Ultrahost/ docs
All examples

Example 09

Experiment

Semantic Data Modeling

Turn what an expert knows about a domain or process into a model of its things, terms, and relationships.

Before data can be connected, it has to be described. This example defines what each document contains, how documents relate to each other, and which terms mean the same thing across all of them, then brings that model to life in a graph.

Chapter one

When an expert describes their world

The people who know a domain best do not have to write code. They decide what exists, how it connects, and what things are called, and the Semantic engine holds those decisions.

  1. 1
    Schemas & Ontology Versioning

    Each kind of document gets a shape

    Which fields matter, which are required, how many of something is allowed, what the defaults are, and what each thing is named.

  2. 2
    Schemas & Ontology Versioning

    Documents are related through an ontology

    The ontology says what connects to what, and just as importantly, what should never connect.

  3. 3
    Schemas & Ontology Versioning

    A dictionary settles the vocabulary

    Shared terms are referenced across documents, so the same thing has the same name everywhere and nothing is left ambiguous.

Chapter two

When the model comes to life

A model on paper can only be read. Once it is instantiated, it can be queried and walked.

  1. 1
    Graph Database

    The ontology is instantiated

    The Graph engine turns the described entities and relationships into a real graph, following the rules the expert set.

Make it yours

Variations worth trying

  • Model one document type completely before relating it to the others.
  • Start with the dictionary alone if your main problem is inconsistent naming.
  • Leave the graph out until you need to query the relationships, not just define them.

Starting point

Ask your coding agent for your version

Use this prompt as a starting point, then change whatever does not fit your idea.

Build a semantic data model with typed documents, ontologies that relate them, shared dictionaries that remove ambiguity, and a graph that instantiates the ontology.

Use these Ultrahost building blocks, in this flow.

When an expert describes their world:
1. Schemas & Ontology Versioning: Which fields matter, which are required, how many of something is allowed, what the defaults are, and what each thing is named.
2. Schemas & Ontology Versioning: The ontology says what connects to what, and just as importantly, what should never connect.
3. Schemas & Ontology Versioning: Shared terms are referenced across documents, so the same thing has the same name everywhere and nothing is left ambiguous.

When the model comes to life:
1. Graph Database: The Graph engine turns the described entities and relationships into a real graph, following the rules the expert set.

Treat this as a starting point. Keep the parts that fit my idea and adapt the rest.

Next example

AI Model Gateway