n_sample -> N

[hbaniecki]

Following ModelOriented/DALEX#175 I want to deprecate n_samples.
Additionally, @pbiecek why is this sampling using replace?

ingredients/R/feature_importance.R

Line 196 in 9f8f82f

sampled_rows <- sample.int(nrow(data), n_sample, replace = TRUE)

No such behaviour here:

ingredients/R/partial_dependence.R

Line 109 in 9f8f82f

ndata <- data[sample(1:nrow(data), N),]

[pbiecek]

replace = TRUE means that nrow(data) can be smaller than n_sample (because some observations ill be reused).
On average it shall not make any difference if we use sampling with/without replacement.