This artifact provides the implementation of (a blue-fringe version of) the "regular positive negative inference" (RPNI) learning algorithm as presented in the paper "Inferring regular languages in polynomial update time" (https://doi.org/10.1142/9789812797902_0004) by Jose Oncina and Pedro García, including merging heuristics such as the "evidence-driven state merging" (EDSM) and "minimum description length" (MDL) strategies. More details on these ...