PyPhi: A toolbox for integrated information theory
Fig 1
(A) The SystemIrreducibilityAnalysis object is the main output of the software. It represents the results of the analysis of the system in question. It has several attributes (grey boxes): ‘ces’ is a CauseEffectStructure object containing all of the system’s Concepts; ‘cut’ is a Cut object that represents the minimum-information partition (MIP) of the system (the partition of the system that makes the least difference to its CES); ‘partitioned_ces’ is the CauseEffectStructure of Concepts specified by the system after applying the MIP; and ‘phi’ is the Φ value, which measures the difference between the unpartitioned and partitioned CES. (B) A Concept represents the maximally-irreducible cause (MIC) and maximally-irreducible effect (MIE) of a mechanism in a state. The ‘mechanism’ attribute contains the indices of the mechanism elements. The ‘cause’ and ‘effect’ attributes contain MaximallyIrreducibleCause and MaximallyIrreducibleEffect objects that describe the mechanism’s MIC and MIE, respectively; each of these contains a purview, repertoire, MIP, partitioned repertoire, and φ value. The ‘phi’ attribute contains the concept’s φ value, which is the minimum of the φ values of the MIC and MIE.