Releases: JuliaApproximation/BasisFunctions.jl
Releases · JuliaApproximation/BasisFunctions.jl
v0.7.8
v0.7.7
v0.7.6
BasisFunctions v0.7.6
- DomainSets is updated to v0.7.18, in preparation for accommodating DomainSets both in the v0.7 and v0.8 series
v0.7.5
BasisFunctions v0.7.5
v0.7.4
BasisFunctions v0.7.4
v0.7.3
BasisFunctions v0.7.3
v0.7.2
v0.7.1
BasisFunctions v0.7.1
This is a non-breaking release. New features:
- the inverse of a function can be computed in a number of limited cases (monotonic function on an interval)
- Gram Schmidt orthogonalisation of a basis was added (using the modified Gram Schmidt algorithm)
v0.7.0
BasisFunctions v0.7.0
Breaking changes
- an expansion (dictionary+coefficients) no longer acts as the vector of coefficients. This means it cannot be indexed (as in
e[2]returning the second coefficient) and cannot be iterated over - there are some interface cleanups. The
normandmomentfunctions with a combination of dictionary and coefficient as arguments are renamed todict_normanddict_momentinstead. One can still invokenormon a basisfunction, e.g.,norm(dict[3])
Features
- computing with expansions has improved. For example, computing the diagonal of the Christoffel-Darboux kernel of a polynomial basis can be done as follows:
julia> basis = orthogonalize(Legendre(3))
DiagonalOperator(v) * Legendre(3)
v = 3-element Vector{Float64}:
0.707107
1.22474
1.58114
julia> k = sum(phi^2 for phi in basis)
Expansion(Legendre(5), v)
v = 5-element Vector{Float64}:
1.5
8.28036e-16
1.71429
1.14325e-15
1.28571v0.6.5
BasisFunctions v0.6.5
- supporting routines for the Plots package have been moved to an extension conditional on loading RecipesBase
- similarly, support for PGFPlotsX has moved to an extension
Closed issues:
- FillArrays compat very tight (#56)