ComplexComputation<T> class abstract base

A versatile custom node computation for handling multiple inbound dependencies (N:1).

ComplexComputation manages an N:1 set of upstream dependencies declared via manifest. It automatically coordinates dirty propagation and lazy reverse-traversal computation.

Design Philosophy: Complex business logic often requires aggregating multiple data sources. ComplexComputation decouples dependency registration from calculation: manifest declares the graph connections while compute computes the combined result lazily.

Data Flow Architecture (Push-Dirty, Pull-Data):

• Push-Dirty (Downstream Propagation): When any upstream node declared in manifest becomes dirty, ComplexComputation marks itself as dirty and forwards the signal without executing compute.
• Pull-Data & Lazy Computation (Reverse Traversal): When a downstream terminal requests data, compute is invoked once per computation cycle, caching the result.

AI & Developer Note:

• Pure Function Constraint: The compute method MUST be pure and synchronous. Do not perform side-effects (e.g., UI updates, network calls) inside compute.
• Manifest Alignment: Every node accessed inside compute MUST be returned in manifest.

Example:

class SumComputation extends ComplexComputation<int> {
  SumComputation(this.a, this.b);
  final Coral<int> a;
  final Coral<int> b;

  @override
  @manifestSync
  Iterable<CoralNode> manifest() => [a, b];

  @override
  int compute() => a.data + b.data;
}