bug: prof.to.features fails when peak shape is Gaussian

[maximskorik]

Description

prof.to.features fails when executed on mbr_test0_cdf.Rds data with shape.model=Gaussian.

The error occurs at the runtime of normix() function which tries to access a value out of range during iteration:

for (m in 1:l) {
   w[m, ] <- dnorm(x, mean = miu[m], sd = sigma[m]) * sc[m] / all.w
}

There are cases when l(lowercase L) is 0. In such cases iteration goes from 1 to 0 but the indexing in R starts at 1.

Steps:

• run the prof.to.features function with the peak shape parameter set to Gaussian by adding a test case which reproduces this error
• use different datasets from the input to see which behaviours occur

[hechth]

Lets first have a look and try some simpler solutions before spending too much time on it.