Lenses

DNA's nouns (resource types) and its connections (relationship types) describe the graph. A Lens names a pattern over that graph: a small set of typed node slots, the directed edges between them, and an optional sentence that reads the pattern in plain language.

The same Lens definition works in two directions:

• Query — find every subgraph in a DNA document that matches the pattern (e.g. every subject holds assignment within boundary binding that exists).
• Command — assert a specific binding of the pattern into the graph (e.g. grant this subject that assignment). No separate schema is needed for each direction.

The LensType contract

Every lens — the core ones below and any you declare inside a DNA document's lenses block — composes a shared base contract:

• name — a human-readable name for the pattern.
• nodes[] — the typed slots. Each declares a type (the resource type it holds) and, when an edge or the sentence references it, a slot identifier.
• edges[] — directed connections between slots. Each declares from and to (slot names) and via (the relationship type traversed). Layer lenses omit edges entirely.
• sentence — an optional template with {{slot}} markers, one per declared slot, giving a readable rendering of the pattern.

The lenses below ship with @dna-codes/dna-core and fall into two shapes: layer lenses, which group a whole layer's resource types with no edges, and subgraph & traversal lenses, which wire a few slots together along named relationships.

Layer lenses

Each layer lens groups all the resource types in one DNA layer — nodes, no edges.

Subgraph & traversal lenses

These lenses connect resource types across relationships, with a sentence that reads the pattern aloud.