pygplib 2.2.0

v2.2.0

fixed bug

pygplib 2.1.0

• Made it possible for constructor of Cnf to accept not only tuple but also list as input.

pygplib 2.0.5

• propnize() and reduce() have been deprecated and will be removed in v3.0.0.
• Renamed propnize() to perform_boolean_encoding().
• Renamed reduce() to reduce_formula().
• Renamed bipartite_order to partitioning_order.
• Cleaned code
• Updated documentation

pygplib 2.0.3

• Added log-encoding

pygplib 2.0.2

• Fixed direct-encoding of GrSt class
• Fixed initialization of Ecc object for a graph with isolated vertex (vertex without neighboring vertex).

pygplib 2.0.1

• Improved documentation.
• Implemented the code for the edge-clique-cover computation.
• Implemented direct-encoding for domain of discourse.
• Removed tools/ and pygplib/util.py to remove dependencies to third-party tools.
• Renamed Fo class to Fog class.
• Updated GrSt class so that
  • object is initialized with a vertex-list and a edge-list, and the ECC computation is performed during the initialization.
  • the methods for the interpretation of relation symbols = and edg as well as compute_domain_constraint() are provided by GrSt, making Fog class and op module independent of domain encoding.
• Changed the format of first-order formulas (negation, existential and universal quantifiers) to make it compatible with TPTP format.
• Renamed decode_assign() of Cnf to decode_assignment() and changed an output assignment so that auxiliary variables are ignored.
• Renamed get_interpretation_of_assign() of GrSt to decode_assignment() and changed an output assignment so that first-order variables' indices are associated with constant symbol indices.
• Removed field st in formula class and changed to give relational structure as argument of each method that requires it.
• Changed API of compute_domain_constraint() so that the input is an index of a free variable and the output is a Prop formula object.

First Release

v1.0.0

Delete CHANGES