the CZT package was completely rewritten. Now supporting destination sizes and plans

[RainerHeintzmann]

There are two things to discuss before merging:

1. Look at this line:

using FourierTools, NDTools, Test
y = randn(ComplexF32, (5,6))
zoom = (2.0,2.0)
@test ≈(real.(czt(y,zoom)),  select_region(upsample2(ft(y), fix_center=true), new_size=size(y)), rtol=1e-5)

This would normally be expected to be equal. Yet a closer inspection reveals, that an even sized array which is boadering to infinitely many zeros, into who's fouriertransform you zoom, should actually yield a different result. This also has to do with the position in the source, which is regarded as the center.

2. It may be useful to use a named argument (e.g. new_size) instead of the unnamed dst_size, since this would require spelling out the dimensions even if one wants to use the default.

[codecov]

Codecov Report

Merging #34 (21dec92) into main (de6204a) will decrease coverage by 0.02%.
The diff coverage is 94.89%.

@@            Coverage Diff             @@
##             main      #34      +/-   ##
==========================================
- Coverage   93.01%   92.99%   -0.02%     
==========================================
  Files          17       17              
  Lines         959     1028      +69     
==========================================
+ Hits          892      956      +64     
- Misses         67       72       +5     

Impacted Files	Coverage Δ
src/czt.jl	93.51% <94.89%> (-2.56%)	⬇️

... and 1 file with indirect coverage changes

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[roflmaostc]

Ready to merge?