How to select ROI for raman notebook example?

[Matheus-gif]

I'm trying to replicate the shared notebooks, where one fits the Raman spectra of glass, but I'm not managing to correctly subtract the baseline. Since I'm not able to use the "gcvspline" method for baseline(), I opted for "als" instead. Here's my attempt:

using JuMP, PyPlot, Ipopt, Spectra, DataFramesMeta, Statistics, CSV

path = raw"C:\Users\User\Desktop\tratar_dados\codigos_julia\doutorado\dems\LS4.txt"
df = CSV.read(path, DataFrame)

x = df[!, 1]
y = df[!, 2]

figure()

plot(x, y, color="black")

xlabel(L"Raman shift, cm", fontsize = 14)
ylabel("Normalized intensity, a. u.", fontsize = 14)
title("Figure 1: the spectrum of interest",fontsize = 14, fontweight = "bold")

display(gcf())

# the regions of interest roi
roi = [860.0 870.0; 1300.0 1400.0]

# evaluate baseline
y_corr, y_bas = baseline(x, y, roi, "als", p=0.01, lam=10^6, niter=10) 

#Creates a plot showing the baseline
figure()
plot(x, y_corr, color="black",label="Treated")
plot(x, y_bas, color="blue",label="Baseline")
plot(x, y, color="green",label="Raw data")

title("Figure 2: the fit of the background",fontsize = 18, fontweight = "bold")

# we set the values of the labels of the x and y axis.
xlabel(L"Raman shift, cm",fontsize=18) # The L in front of the string indicates that we use Latex for the maths
ylabel("Intensity, a. u.",fontsize=18)

# we display the legend at the best location, without a frame
legend(loc="best",frameon=false)

display(gcf())

[charlesll]

ALS does not use ROI because it is supposed to be an automatic algorithm that actually does not require ROIs. This is why it does not work.

Only the gcvspline, unispline and poly use the ROI.

• I need to update the docs to clarify this!

[Matheus-gif]

hmm, that is nice! It looks like I'm forced to input a ROI as an argument though.
I've restarted the kernel and run the following, leading to a MethodError:

# the regions of interest roi
roi = [860.0 870.0; 1300.0 1400.0]

# evaluate baseline
#y_corr, y_bas = baseline(x, y, roi, "arPLS", p=0.01, lam=10^6, niter=10) 
y_corr, y_bas = baseline(x, y, "als", p=0.01, lam=10^6, niter=10) 

#Creates a plot showing the baseline
figure()
plot(x, y_corr, color="black",label="Treated")
plot(x, y_bas, color="blue",label="Baseline")
plot(x, y, color="green",label="Raw data")

title("Figure 2: the fit of the background",fontsize = 18, fontweight = "bold")

# we set the values of the labels of the x and y axis.
xlabel(L"Raman shift, cm",fontsize=18) # The L in front of the string indicates that we use Latex for the maths
ylabel("Intensity, a. u.",fontsize=18)

# we display the legend at the best location, without a frame
legend(loc="best",frameon=false)

display(gcf())

MethodError: no method matching baseline(::Vector{Float64}, ::Vector{Float64}, ::String; p::Float64, lam::Int64, niter::Int64)
The function `baseline` exists, but no method is defined for this combination of argument types.

Closest candidates are:
  baseline(::Array{Float64}, ::Array{Float64}, !Matched::Array{Float64}, !Matched::AbstractString; polynomial_order, s, lam, p, ratio, niter, p0_exp, p0_log)
   @ Spectra C:\Users\User\.julia\packages\Spectra\gifVq\src\baseline.jl:85


Stacktrace:
 [1] top-level scope
   @ c:\Users\User\Desktop\tratar_dados\codigos_julia\doutorado\dems\jl_notebook_cell_df34fa98e69747e1a8f8a730347b8e2f_W2sZmlsZQ==.jl:6

[charlesll]

Yes, this is normal and written in the documentation of the function: "Note: This is NOT used by the "als" and "arPLS" algorithms, but still is a requirement when calling the function. bir and method probably will become args in a futur iteration of rampy to solve this."

Note that there is also a mismatch in the docstrings, with bir = roi.

I need to solve that by setting roi as optional arguments, but I did not get the time to do it yet.

Just pass anything as ROI and it will work, the algo does not consider it as long as you use als and arPLS.

[charlesll]

This is going to be fixed by the next version, working on it right now.

[Matheus-gif]

Got it, thank you for your assistance :)

[charlesll]

@Matheus-gif Version 2.0.1 solves this issue, check the new doc and the new code!