aGrUM/pyAgrum 1.1.0 released

ANNOUNCE: aGrUM 1.1.0

aGrUM/pyAgrum 1.1.0 is out.

Changelog for 1.1.0

• aGrUM

  • Parallelizing inference for Markov Net : gum::ShaferShenoyMNInference.
  • Adding a new gum::fastVar(std::string) factory function to easily build variable such as in gum::*::fastPrototype(std::string). See Documentation. This syntax is now used in every graphical model with gum::{BayesNet|MarkovNet|etc}::add(std::string) to easily add a new variable in the model.
  • Adding a new graphical representation of cliqueGraph : gum::CliqueGraph::mapToDot().

• pyAgrum

  • Following aGrUM, gum.ShaferShenoyMNInference is now parallelized too.
  • Wrapping the new fast syntax for creating variable in functions gum.{BayesNet|etc}.add(str). See Documentation.
  • Introducing a better type induction for discrete variables when created by gum.skbn.BNDiscretizer. gum.IntegerVariable and gum.RangeVariable are now induced from the data.
  • Adding gum.JunctionTree.map() (in notebook) to show a condensed map of a the junction tree.
  • Still working on documentation (One must imagine Sisyphus happy).