exalearn / ExaMol / 5601334459

98%

push

github

Add support for multi-objective optimization (#77)

* Start by changing the specification to use >1 recipe

We're going to be multi-objective by default

* Make the SimulationRequest hashable

Do so by making it frozen

* Port the steering class to support more than one object

We should rename it next

* Continue updating steering algorithm for multiobjective

More issues around properly accounting for a 2D list of models

* Update score functions to account for >1 objective

* Update the interface for selectors

* Have reporters work with >1 recipe

Also fix #72 while at it

* Exit immediately after Thinker is done

* Update specification for example

* Rename the thinker

We can now do multi-objective optimization

* Re-add files needed by tests

* Improve test coverage

And mark error messages as not coveraged

* Fix BO test

We needed to adjust to the new test fixture

111 of 111 new or added lines in 9 files covered. (100.0%)

1627 of 1652 relevant lines covered (98.49%)

1.97 hits per line