HMFW() a function to compute the "half-maximum full-width" of peaks in spectra

[aphalo]

This function is rather easy to implement with existing functions. A first cut at the code for a single peak is below. This could be made an S3 method for generic_spct but probably not needed, while support for multiple peaks in a single spectrum much more important.

HMFW <-
  function(
    x,
    span = NULL,
    ignore_threshold = 0,
    strict = TRUE,
    na.rm = FALSE,
    refine.wl = FALSE,
    method = "spline",
    ...) {
    stopifnot(is.any_spct(x))
    peaks.found <- peaks(x = x,
                         ignore_threshold = ignore_threshold,
                         span = span, 
                         strict = strict,
                         na.rm = na.rm)
    if (nrow(peaks.found) == 0) {
      return(NA_real_)
    } else if (nrow(peaks.found) == 1) {
      wavelengths <- wls_at_target(x = x,
                                   target = "0.5max",
                                   interpolate = refine.wl,
                                   na.rm = na.rm)[["w.length"]]
      if (length(wavelengths) == 2) {
        return(diff(wavelengths))
      } else {
        return(NA_real_)
      }
    } else {
      message("Not implemented yet for multiple peaks")
    }
  }

• Add function
• Add unit tests
• Document help page
• Mention in User Guide

[aphalo]

While wls_at_target() does the job of finding HM values these are not referenced to the nearest peaks. It could be better to make the finding of the HMFW wavelengths and widths and option in peaks() rather than creating a new function. The HMFW of a peak must be computed relative to each peak's maximum, not the global maximum as done by wls_at_target().

So this is not as simple as using the code chunk above. So, moved to a future Milestone. It has to be implemented after or at the same time as #29 and #30 are addressed.