v0.5.0

This is a breaking API release, where #328 had huge impact on the use of the library.
The PR details the main breaking changes that concerns the name of packages, modules, class. Everything is now realigned to a convention
As described in the PR :

• We use camel case for classes
• For each problem,
  • the solvers are stored in a subpackage "solvers" (was sometimes "solver"),
    in a module with a short name "{technology}" without adding
    "solver" or the problem name
  • the base solver class is also in solvers, accessible directly via
    discrete_optimization.{problem}.solvers,
  • the solvers_map is accessible in discrete_optimization.{problem}.solvers_map,
  • the solver class name follow the logic {Technology}{Problem}Solver,
  • the problem class name is {Problem}Problem,
  • the solution class name is {Problem}Solution,

Raise issue if any need for adaptation of code or undesired effect.

Main new adding : didppy usage on most problems implemented. We thank the developers of this lib for the support.

What's Changed

• Adapt pickup-vrp lp iterative solver with ortools/mathopt by @nhuet in #320
• Fix wrong implementation of the capacity check of the facility problem. by @g-poveda in #322
• Minor changes in fjsp solvers by @g-poveda in #323
• Fix vrp callback in solver ortools by @g-poveda in #324
• New solvers based on dynamic programming and the didppy library by @g-poveda in #321
• Adding example on solving rcpsp with calendar by @g-poveda in #325
• Bug fix by @g-poveda in #326
• Full renaming (package/module/class) by @nhuet in #328
• Correct rcpsp solvers_map module name by @nhuet in #329
• Avoid warning at each iteration of lns about subsolver time_limit by @nhuet in #331
• Github runner macos-12 replaced by macos-13 by @nhuet in #330
• TSP with ortools update by @g-poveda in #332
• Implement warmstart for DP solver/models by @g-poveda in #333
• Avoid module names which match a known library name by @nhuet in #334
• Rename test_ortools into test_ortools_routing by @nhuet in #336
• Test use of callbacks with dp solvers by @nhuet in #337
• Detect by introspection which arguments are licit for dp solvers by @nhuet in #338
• Fix lns init phase + constraint handler mix for minizinc subsolvers by @nhuet in #335
• Readme update by @g-poveda in #340
• Update test with callback for rcpsp/cp_mzn that randomly failed by @nhuet in #341
• Use an internal indexing of nodes to allow CPM to work with mixed nodes by @nhuet in #339
• Handle failed introspection on the dp solvers by @g-poveda in #342

Full Changelog: v0.4.0...v0.5.0