This version of the DYCI2 library is the first release of a work in progress: a set of python modules defining models and tools for creative generation of sequences (and in particular musical sequences) from models of sequences. It implements several models, generative heuristics, time management strategies, and architectures of interactive agents. Each class is associated with a tutorial (see the tutorials in the root directory).
DYCI2 ("Creative Dynamics of Improvised Interaction") is a collaborative research and development project funded by the French National Research Agency (ANR). It explores the creative dynamics of improvised interactions between human and artificial agents, featuring an informed artificial listening scheme, a musical structure discovery and learning scheme, and a generalized interaction / knowledge / decision dynamics scheme (see http://repmus.ircam.fr/dyci2/home). The DYCI2 library is part of the DYCI2 project, it is conceived as an autonomous and easily extensible Python library, and can also be used in association with audio or midi listeners and renderers to form DYCI2 agents (see directory "Max").
More information on the project: refer to Nika, Déguernel, Chemla--Romeu-Santos, Vincent, Assayag, "DYCI2 agents: merging the "free", "reactive", and "scenario-based" music generation paradigms", in Proceedings of International Computer Music Conference 2017 (https://hal.archives-ouvertes.fr/hal-01583089/document).