NEWS

ir 0.4.2 (2026-02-03)

ir_scale(): Function to scale intensity values in spectra, similar to base::scale(). Intensity values are scaled for the same x axis values.
ir_as_ir.ir_flat(): Conversion of ir_flat objects to ir objects.
ir_as_ir.hyperSpec(): Conversion of hyperSpec::hyperSpec objects to ir objects.
ir_as_ir.Spectra(): Conversion of ChemoSpec::Spectra objects to ir objects.
ir_identify_empty_spectra(): Identifies rows in an ir object with empty spectra (former internal function ir_check_for_empty_spectra()).
ir_sample_prospectr(): Wrapper function to use sampling algorithms implemented in the 'prospectr' package directly with 'ir' objects.

ir_correct_atmosphere() has a new parameter return_contribution which allows to return the contribution of ref to each spectrum in x.
ir_normalize() has three new methods to normalize spectra:
1. With method = "area_absolute", all intensity values will be divided by the sum of the absolute intensity values at all wavenumber values of the spectrum.
2. With method = "vector", all intensity values will be divided by the norm of the intensity vector (vector normalization).
3. With method = "snv", a Standard Normal Variate correction will be performed.
ir_bc_rubberband() has an additional parameter to allow imputation of the first and last value in a spectra to avoid artifacts which may happen during baseline correction with the rubberband algorithm.
All functions now fully support empty spectra.
Several dependencies in Imports have been moved to Suggests.
Efficiency of several functions has been improved.

In ir_import_spc(), the returned ir object was not a tbl_df, but a data frame. This is now corrected.
In unnest.ir(), the ir class was not restored and therefore the result was always unclassed.
In ir_average(), the returned ir object was still a grouped_df. This is now corrected (by adding the .groups argument for dplyr::summarize() which is internally used by ir_average()).

ir_export_prepare() to support exporting ir objects to csv.
ir_correct_atmosphere() enables correcting for atmospheric artifacts (e.g. H$_2$O and CO$_2$ peaks) following Perez-Guaita et al. (2013) (DOI: 10.1366/13-07159).

Making ir_flatten() faster by improving spectra combining.
Making ir_bin() faster by improving spectra summarizing per bin. Allowing to specify how wavenumber values should be defined after binning.
Arithmetic operations now also accept numeric vectors (ir_subtract(), ir_add(), ir_multiply(), ir_divide()).

Circumventing a bug in ir_import_spc() by allowing to not export spectra metadata. This bug is caused by hyperSpec::read.spc() and will be fixed in the upcoming weeks. See https://github.com/r-hyperspec/hyperSpec/issues/80

First CRAN release (#1).
Added a NEWS.md file to track changes to the package.
Added tidyverse methods for ir objects.
ir objects no longer require columns measurement_id and sample_id, but only the column spectra.
Added subsetting and binding methods for ir objects.
Added arithmetic operators for ir objects.