COGANT: Deterministic Codebase-to-GNN Translation

COGANT (Codebase-to-GNN Translation) deterministically converts software repositories into structured Active Inference artifacts expressed in the Active Inference Institute's Generalized Notation Notation (GNN — Generalized Notation Notation , not graph neural networks). It is best understood as an evidence compiler : it propagates reviewable program facts through a fixpoint rule pipeline (in the tradition of fixpoint program analysis) and emits graph, matrix, provenance, visualization, and roundtrip artifacts rather than a single opaque embedding or proof. A program-graph intermediate representation (IR) carries confidence and provenance; a fixpoint translation engine applies 22 declarative rules mapping nodes onto 7 Active Inference mapping kinds . Each emitted bundle includes a seed-induced structural partition using Markov-blanket role vocabulary, and a reverse synthesizer reconstruc