Release v0.10.1

Release notes

Fixed

This is a hotfix release due to a bug discovered in bounds computation.

• Bounds for pow and mod (cherry picked for this hotfix) #1012

Full Changelog: v0.10.0...v0.10.1

Release v0.10.0

Release notes

Added

• New solver: Rc2 MaxSAT solver #729
• Expended to_cnf using pindakaas encoding backend #782
• Linearisation of multiplication between Boolean and Integer #769
• Solution callback for Hexaly #787, #809
• Start of solver parametrised testsuite #817
• Documentation overview of solver capabilities #728
• Typehints and documentation update for global constraints #812
• Typehints for Solve and SolveAll #775
• OCUS in tools, allowing meta-constraints on MUSes #698
• Fix supported solver version ranges #816
• Transversal / hitting set example (contributed by @nandorsieben) #790
• Nurserostering benchmark #789
• Prediction+optimisation example of scheduling surgeries under uncertainty 5a87c5e
• Decision-focused learning example #621

Internal improvements

• Globals define their own negation #703
• Change cp.sum(*iterable, **kwargs) to cp.sum(iterable, **kwargs) #756
• Division and Modulo as global functions #807
• Refactor decompose for global functions #793
• Refactor and update Cumulative and NoOverlap constraints #694

Changed

• Throw ModuleNotFoundError when module is not installed #825

Fixed

• Fix support for newest OR-Tools 9.15 release #821
• Check Hexaly license before using solver #826
• Fix to_cnf clause bypass issue #824
• Fix bug handle pdk unsat with conditions #811
• Missing packaging dependency in setup.py #813
• Abs constraint handle None #794
• Timeout under assumptions for Exact #805
• Support for expressions in start, duration and end in CPO #802
• Consistent version checks #792
• Missing constraint tags for Pumpkin #799
• Z3 negate maximisation objective #786
• Scaled Booleans in Pumpkin interface #776

Full Changelog: v0.9.29...v0.10.0

v0.9.29

Release notes

This is a hotfix release due to external breaking changes in the default solver backend.

Fixed

• Pin solver versions (not merged, cherry picked for this hotfix) #816

Full Changelog: v0.9.28...v0.9.29

v0.9.28

Release notes

This is a very small release with a hotfix for multi-dimensional indexing where the index is a decision variable.

Fixed

• Hotfix for single expression in multi-dimensional indexing #772

Full Changelog: v0.9.27...v0.9.28

v0.9.27

Release notes

Added

• New solver: CPLEX ILP solver #583, #762, #763
• New solver: Hexaly local search solver #718
• New tool: CPMpy cli (get version information on all subsolvers) #693
• Use of CSE in flatten_objective #730
• Constraint tagging for pumpkin #720
• Solver capability overview in docs d8d3a2d
• PSP-lib dataset and loader #701
• Safen name collision between user and aux variables #731

Changed

• Update pindakaas to version 0.2.1 #753
• Exclude slow examples from testset #746
• Safe variable names #731
• SolverInterface consistency improvements and added tests #726
• Improve testing of examples #651
• Modernize shebangs of examples #644
• Linearize improvements e721b0a
• Update waterfall (added automated int-to-bool) 710ec42
• Collection of improvements to GitHub README

Fixed

• Precedence global decomposition #742
• Minor bugs in examples #739
• gnureadline typo in setup.py #722
• Non-contiguous array handling in cpm_array #738
• Remove simplification of nullifying arguments from constructors #725
• Skip psplib in testset due to random failure on GitHub actions 7a15d07
• (Pin Pumpkin solver git commit #719) outdated by #720

Full Changelog: v0.9.26...v0.9.27

v0.9.26

Release notes

Added

• New solver: Pumpkin LCG solver (by the ConSol Lab at TU Delft) #669
• New solver: Pindakaas library for transforming pseudo Boolean to CNF #600
• New global: Regular #677
• New tool: Tooling for the XCSP3 format: dataset, loader, CLI, benchmarking #597
• Common Subexpression Elimination (CSE) #679
• Support for integer constraints for Boolean solvers using new int2bool transformation #653
• Solution hinting for Gurobi #691
• Solution callback for CP Optimizer #682
• Acces to (installed) solver status/version: call .version() on a specific solver or on SolverLookup to get an overview. Or use cp.SolverLookup.print_version() for pretty printing. #628
• Support for canonicalisation of subtraction expressions #686
• Allow use of Boolean constants in weighted sum #711
• Added testset for incremental assumptions
• Catch beginner mistakes linked to incorrect usage of boolean expression #660

Changed

• Refactoring of int2bool/pysat #714
• Updated instructions for adding a new solver with request to add solver to overview table
• Improve efficiency only_implies: avoid retransform if no subexpression #680
• Consistent state reporting of .solve() across all solvers #545
• Improved docs and error messages for canonical_comparison #678

Fixed

• Support special cases for Xor global (boolean constants, single argument) #717
• Missing optimal exit status for GCS #705
• Missing solution for empty csp when enumerating #674
• Use of constants in globals #700
• CSE edge-case in only_numexpr_equality #696
• Choco shorttable star symbol data type #697
• Ensure python-native expression values to prevent unexpected behaviour with numpy constants #695
• Correctly handle numpy integers in eval_comparison #683
• CP Optimizer fix cumulative with zero-duration task #681
• Fix incorrect decomposition of Inverse global constraint #673

Full Changelog: v0.9.25...v0.9.26

v0.9.25

Release notes

Added

• New solver: IBM CP Optimizer #576
• New global: ShortTable #469
• model.add() as new default to add constraints to model (same behaviour as +=) #640
• Easy install of all solvers through the "all" optional dependency, i.e. pip install cpmpy[all] #665
• More complex variants of the Sudoku puzzle (in examples) #577
• API summery sheet #629
• Linearisation for div, abs and mod #516
• Linearisation for subtraction #639
• div linearisation for Gurobi #593
• Native cardinality constraints for PySAT #588
• wsum support for PySAT
• Support for constants in Element global function #630
• SolverLookup now has .supported() and .print_status() to get information on the available solvers on the current system #641
• solveAll() now accepts solver-specific kwargs, just like solve() #582
• Optional dependencies for solvers and tools (setup.py) #599
• Documentation for all CPMpy exceptions #622

Changed

• Bumped minimal Python version from 3.7 to 3.8 #575
• mod and div now default to integer division
• Improved reading of DIMACS formatted files #587
• Avoid max-based decomposition for abs if possible #595
• Cleanup semantics of Boolean Python builtins #602
• Performance optimisation of simplify_bool #592
• Ensure AllDifferent decomposition is linear #614
• Much improved README
• Improved docs formatting, especially for the Python API docstrings #603
• Improved API documentation of explanation tools #512
• Better exceptions for explanation tools #512
• Improve documentation of non-standard solver args passing for Exact #616
• General documentation improvements #619, #633, #634
• Skip subset of PySAT tests when optional dependency pblib is not available #668

Fixed

• Linearisation with boolean constants #581
• Linearisation of AllDifferent with integer constants #547
• Side conditions for Precedence global #589
• simplify_bool on non-CPMpy expressions #626
• Handling of negative variables in objective during linearisation #495
• Integers in GCC global for Choco-solver #646
• NValueExceptN for single value range #645
• Handling of empty clauses in GCS #662
• Missing user vars when calling solveAll(), resulting in incorrect number of solutions #609
• Consistent handling of non-positive time_limit #642
• Added setuptools to required dependencies (can be missing on some installs) #664
• DIMACS tempfiles on Windows #601

Removed

• Removed support for Python version 3.7 #575

Full Changelog: v0.9.24...v0.9.25

0.9.24

Release Notes

Enhancements and Features

• Safening of partial functions: New transformation to safen partial function. #515
• Update Choco interface with native reification: Improved the Choco solver interface with native reification support. #526
• Reals in objective function: Added support for real numbers in the objective function. #529
• Better naive grow: Enhanced the "naive grow" strategy for better performance. #528
• Blocks world: Introduced a "blocks world" example for demonstration purposes. #533
• Examples Colab links: Added direct Colab links to examples for easier experimentation. #553
• Circuit decomposition for all ranges: Extended circuit decomposition to handle all ranges. #424
• Global function in Z3 objective: Introduced global functions in Z3 solver objectives. #560
• Z3 auto subsolver: Implemented automatic subsolver selection for Z3 when solving optimization problems. #567
• Parametrize solver tests: Streamlined solver test cases using parameterization. #572
• Linearize power: Added linearization for power operations. #538
• Improve Boolean normalization for PySAT: Enhanced normalization of Boolean terms for PySAT. #569

Performance Improvements

• Has subexpr optimization: Skip transformations of leaf expressions for improved efficiency. #532
• Only implied speedup: Optimized "only implied" handling for significant speedups. #541
• Distribute tests over CPUs: Distributed tests over 4 CPUs, reducing runtime from 21m30s to 8m45s. #571
• ndvar_getitem optimization: Improved ndvar_getitem by moving fetch casts to initialization. #550
• Remove inline imports: Instead use import as x at top of file. #542

Bug Fixes

• Remove broadcast in min/max: Fixed issues when forwarding to built-in min/max functions. #536
• Convert numpy array in Table constraint: Ensured proper conversion of NumPy arrays to lists in Table constraints. #540
• Clear values on UNSAT: Added functionality to clear variable values when UNSAT is detected. #523
• Fix cpm_array with order='F': Resolved issues with cpm_array() when using column-major order. #555
• Car sequencing index fix: Corrected indexing issues in the car sequencing problem. #565

Code Quality and Maintenance

• Improve exception messages: Enhanced clarity of exception messages and removed unused imports. #539
• Edits to the docs: Updated documentation for clarity and completeness. #530
• Gurobi license check: Separated Gurobi license checks into a distinct process. #566
• Standardize solver version checks: Unified approach to checking solver version compatibility. #568
• Update requirements: Upped our minimal python requirement from 3.6 to 3.7. #573

0.9.23

Quick release, because we want the updated tools to be available.

What's Changed

• Extension to tools: MARCO and SMUS
• Added tests for incremental solving and fixed incemental solving with objective in Exact
• Cumulative decomposition fix when capacity was numpy integer.

Full Changelog: v0.9.22...v0.9.23

v0.9.22

What's New

• New solver: GlasgowConstraintSolver (GCS)
• Upgraded to Exact 2
• Minizinc print: easily extract MiniZinc and FlatZinc text from CPMpy model.
• Update TEMPLATE.py to make it clearer how to add new solvers.
• SolverLookup gives clear error message in stead of returning None
• allow kwargs in Model.solve()
• call python builtins for sum, abs, min and max without expressions in the arguments.
• All solvers now have a native_model() function, to allow native solver access.
• It's now possible to name multiple new variables at once, by providing the names in a list.
• Linearize transformation can now rewrite modulo (if multiplication is supported)
• Fix behaviour of "all", "any", "max", "min", "sum", "prod" on higher dimensional NDVarArrays (maintain dimensionality)
• Value function of expressions now always returns a python integer, where it could sometimes be a numpy integer.
• Fixed performance issue where all solver vars where seen as user vars when solving with MiniZinc

Documentation

• Overall improvement of documentation
• update documentation of 'comparison' transformation

New Contributors

Thanks to 2 new contributors!

• @ThomSerg and @sin3000x

Full Changelog: v0.9.21...v0.9.22