Explanations

This section explains how annnet represents a graph, how its state changes, and how graph data moves between annnet and other tools. Follow the sections in order for the main path, or jump directly to a topic.

1. Understand the model

Principles and architecture

• Design principles: the choices that shape the package.
• Package architecture: the main modules and how they fit together.
• Internal representation: canonical storage, indices, and attribute tables.

Identity and representation

• Edge tables, ids, and formats: what an edge identity means and what each format can preserve.
• Incidence representation: how graph structure appears in incidence matrices.

2. Work with graph state

Inspect and change

• Reading the graph: choose the right frame and namespace for a query.
• Adding edges: accepted edge inputs and how they are dispatched.
• Mutation and derived state: what writes maintain and what reads derive.

Layers and context

• Aspects, order, and windows: organize layers and select them as a query.
• Multilayer and multi-aspect graphs: model structure across layer coordinates.
• Slices and views: work with contexts without duplicating the graph.

Values

• Node-layer values: attach, read, and interpret values on node-layer pairs.

3. Connect and persist

Interoperate

• Interoperability: convert graphs while making projections explicit.
• Storage and IO: persist native state and exchange it with other formats.

Track changes

• Tracking changes: use mutation history, snapshots, and diffs.

When you need exact signatures, parameters, attributes, or methods, use the API reference. For runnable examples, use the Notebook Gallery.