v0.1.4
LieGroups v0.1.4
Added
- mention
adjoint_matrixin the transition documentation (#52) - introduce
jacobian_expfor the Jacobian of the exponential function. - Introduce a
AbstractActionActsOnTypeto distinguish, what previously was called “side”,
i.e. whether an action acts on the left (ActionActsOnLeft) or right (ActionActsOnRight) - Introduce
_invand_inv!functions for the inverse operation to work the same way as_composeand_compose!, respectively. - introduce a
LeftMultiplicationGroupActionto represent left multiplication actions on Lie groups. While this is a bit more of a technical
name, it replaces the oldComplexPlanarRotation,QuaternionRotation, andRotationAction, since in theGroupActionthis type is coupled with a group and a manifold anyway. - further methods for the
LeftMultiplicationGroupActionto cover the previous functionality of e.g.RotationTranslationActionOnVector
- Add the Special Galilean Group (#60)
Fixed
- Fixed an issue where internally the product manifold in a
SemidirectProductLieGroupwas accidentally splashed - make
×andProductLieGroupbehave the same way as×andProductManifolddo - Fixed an issue, where
exp!(G, h, g, X)would return a wrong result if the inputgand the outputhare aliased (#63). - Fixed issues with the documentation of
diff_left_composeanddiff_right_composethat were inconsistent - fixed
push_forward_tangentandpull_back_tangentwhich on general Lie groups provided a wrong default. - fixes an allocation bug for
applyanddiff_apply(#52)
Merged pull requests:
- Porting the SpecialGalileanGroup from RoME (#60) (@Affie)
- Test and fix exp! issue from #63 (#64) (@Affie)
- Transfer more elements from Manifolds.jl (#65) (@kellertuer)
- Improve how product and semidirect Lie group constructors act. (#66) (@kellertuer)
- Refactor the semidirect product Lie group to cover all cases from GroupManifolds. (#67) (@Affie)
- Test 8 Semidirect Product combinations, add compose associativity test, and fix compose alias bug (#68) (@Affie)
- Test Semidirect Product against expected
invresults and fix inv/compose combinations (#70) (@Affie)
Closed issues: