- switched to abstract arrays as suggested by #26
- re-definition of types was necessary, taking into account Julian recommendations, see this discourse entry, see this discourse entryhttps://discourse.julialang.org/t/creating-ones-own-matrix-type/33553, and see the docs.
- modified test functions to account, e.g. for static arrays.
- increased code coverage
- examples work with new type hierarchy
- updated dependencies
- new type hierarchy that involves four abstract types:
AbstractMeasure
AbstractOrthoPoly
AbstractQuad
AbstractTensor
- canonical measures and orthogonal polynomials are now subtypes of
AbstractCanonicalMeasure
andAbstractOrthoPoly
- code hygiene at several fronts
- removed default
show
function for types specific toPolyChaos.jl
as this was causing issues with the Atom editor - improved
showpoly
, and addedshowbasis
- updated documentation
- updated package dependencies
- added example for chance-constrained DC optimal power flow to documentation see 004be7c
- computation of moments from PCE coefficients is now compatible with
JuMP
(was restricted toVector{Float64}
before), see see f99030
- implemented show function for orthogonal polynomials (thanks @pfitzseb )
- improved Golub-Welsch algorithm (adapted from
GaussQuadrature.jl
which itself is based on this Fortran code) - improved code coverage, i.e. added tests
- code hygiene
- improved performance of recurrence coefficients, quadrature rules, evaluation of orthogonal polynomials
- added considerable number of tests for recurrence coefficients (see #4) and quadrature rules (see #5)
- added code coverage
- added GNU General Public License v3.0
- removed dependency on
IterTools
- initial version