Fix signalstats implementation when having samples with units

[fhagemann]

In PR #41, code was introduced that fails if the samples have units:

using RadiationDetectorDSP
using RadiationDetectorSignals
using Unitful

wf = RDWaveform((0:100)u"ns", randn(101)u"C")
signalstats(wf, 0u"ns", 100u"ns")

ERROR: DimensionError: 0.0 C and 1.0373625148197236 C^2 are not dimensionally compatible.
Stacktrace:
 [1] _isless
   @ ~/Software/Unitful.jl/src/quantities.jl:250 [inlined]
 [2] isless(x::Quantity{Float64, 𝐈 𝐓, Unitful.FreeUnits{…}}, y::Quantity{Float64, 𝐈^2 𝐓^2, Unitful.FreeUnits{…}})
   @ Unitful ~/Software/Unitful.jl/src/quantities.jl:240
 [3] max(x::Quantity{Float64, 𝐈^2 𝐓^2, Unitful.FreeUnits{…}}, y::Quantity{Float64, 𝐈 𝐓, Unitful.FreeUnits{…}})
   @ Base ./operators.jl:481
 [4] _signalstats_impl(X::StepRange{Quantity{…}, Quantity{…}}, Y::Vector{Quantity{…}}, idxs::UnitRange{Int64})
   @ RadiationDetectorDSP ~/.julia/packages/RadiationDetectorDSP/zR8oQ/src/signalstats.jl:54
 [5] signalstats(input::RDWaveform{…}, start::Quantity{…}, stop::Quantity{…})
   @ RadiationDetectorDSP ~/.julia/packages/RadiationDetectorDSP/zR8oQ/src/signalstats.jl:21
 [6] top-level scope
   @ REPL[8]:1

After this PR, signalstats does not fail and the output looks like this:

(mean = 0.07191460017009416 C, sigma = 0.9393140224212858 C, slope = -0.0040645798613581545 C ns^-1, offset = 0.27514359323800186 C)

[oschulz]

Thanks!

JuliaRegistries/General#124107