resample with dims is broken

[mchitre]

Resample of a 1D array without specifying dims works, but specifying dims=1 breaks it.

MWE:

julia> versioninfo()
Julia Version 1.11.3
Commit d63adeda50d (2025-01-21 19:42 UTC)
Build Info:
  Official https://julialang.org/ release
Platform Info:
  OS: macOS (arm64-apple-darwin24.0.0)
  CPU: 10 × Apple M1 Pro
  WORD_SIZE: 64
  LLVM: libLLVM-16.0.6 (ORCJIT, apple-m1)
Threads: 1 default, 0 interactive, 1 GC (on 8 virtual cores)
(jl_45BPGW) pkg> st DSP
  [717857b8] DSP v0.8.1

julia> using DSP

julia> x = randn(100);

julia> resample(x, 4) |> length
400

julia> resample(x, 4; dims=1) |> length
ERROR: AssertionError: Length of resampled output different from expectation.
Stacktrace:
 [1] checked_resample_output!
   @ ~/.julia/packages/DSP/VbLcO/src/Filters/stream_filt.jl:719 [inlined]
 [2] (::DSP.Filters.var"#58#59"{FIRFilter{DSP.Filters.FIRInterpolator{Float64}, Float64}, Int64, Vector{Float64}, Vector{Float64}, Int64})(v::Vector{Float64})
   @ DSP.Filters ~/.julia/packages/DSP/VbLcO/src/Filters/stream_filt.jl:774
 [3] mapslices(f::DSP.Filters.var"#58#59"{FIRFilter{DSP.Filters.FIRInterpolator{Float64}, Float64}, Int64, Vector{Float64}, Vector{Float64}, Int64}, A::Vector{Float64}; dims::Int64)
   @ Base ./abstractarray.jl:3281
 [4] mapslices
   @ ./abstractarray.jl:3265 [inlined]
 [5] _resample!(x::Vector{Float64}, rate::Int64, sf::FIRFilter{DSP.Filters.FIRInterpolator{Float64}, Float64}; dims::Int64)
   @ DSP.Filters ~/.julia/packages/DSP/VbLcO/src/Filters/stream_filt.jl:769
 [6] _resample!
   @ ~/.julia/packages/DSP/VbLcO/src/Filters/stream_filt.jl:760 [inlined]
 [7] resample(::Vector{Float64}, ::Int64; dims::Int64)
   @ DSP.Filters ~/.julia/packages/DSP/VbLcO/src/Filters/stream_filt.jl:757
 [8] top-level scope
   @ REPL[8]:1

[wheeheee]

Oh no...I believe the underlying reason is that reset!(::FIRInterpolator) was either accidentally deleted or didn't exist. This was exposed in #623 because reset! was used in the implementation of buffered array resampling.
It looks like this can be quickly fixed in a patch. In the meantime, you could try arbitrary resampling with a rate that is a float (4.0), or just use mapslices with resample directly.