Changelog

irpeat 0.3.0

Major changes

• Major restructuring: The models originally included with the ‘irpeat’ package are now stored in a separate data package ‘irpeatmodels’ to reduce the size of the ‘irpeat’ package. The ‘irpeatmodels’ package can be installed from Zenodo.
• Add a new class irp_prediction_domain: This class stores the prediction domain for a model. Methods available are: Conversion from ir and ir_flat objects, plotting, check whether spectra in an ir object are within a prediction domain.
• Rename:
  1. irp_content() to irp_predict().
  2. irp_holocellulose_2() to irp_holocellulose_content_2().
  3. irp_klason_lignin_2() to irp_klason_lignin_content_2().
  4. Output from irp_hkl_hodgkins(): holocellulose_hodgkins –> holocellulose_content_1, klason_lignin_hodgkins –> klason_lignin_content_1.
  5. Output from irp_eac_1(): eac –> eac_1.
  6. Output from irp_edc_1(): edc –> edc_1.
• For existing and new models: Training and testing prediction domains were added and can now be used to check whether the models cannot be used to make reliable predictions.
• By default, functions return Posterior

New functions

• Add irp_preprocess_for(): Allows to extract the spectra after automated preprocessing as they would be used to make predictions.
• Add irp_get_prediction_domain_for(): Allows to extract the prediction domain for a specific model.

New data

• Add sample data (transmission mid-infrared spectra for peat samples from the ‘redoxpeat’ R package): irpeat_sample_data.

New models

• irp_carbon_content_1(), irp_nitrogen_content_1(), irp_oxygen_content_1(), irp_hydrogen_content_1(), irp_phosphorus_content_1(), irp_potassium_content_1(), irp_sulfur_content_1(), irp_calcium_content_1(), irp_silicon_content_1(), irp_titanium_content_1(), irp_d13C_1(), irp_d15N_1(), irp_nosc_1(), irp_dgf0_1(), irp_bulk_density_1(), irp_loss_on_ignition(), irp_O_to_C_1(), irp_H_to_C_1(), irp_C_to_N_1(), irp_volume_fraction_solids_1(), irp_non_macroporosity_1(), irp_macroporosity_1(), irp_saturated_hydraulic_conductivity_1(), irp_specific_heat_capacity_1(), irp_dry_thermal_conductivity_1(), irp_microbial_nitrogen_content_1(), irp_degree_of_decomposition_1(), irp_degree_of_decomposition_2(), irp_degree_of_decomposition_3().

irpeat 0.2.0

• Performance improvement for irp_preprocess().
• Updated irp_preprocess(), model preprocessing configurations, and prediction functions to consider the new arguments (1) new_x_type from ir::ir_bin(), (2) bc_do_impute from ir::ir_bc(), (3) do_return_as_ir (allows to return the preprocessed spectra as ir object).
• Update the documentation: Adding explicit warnings to irp_content_klh_hodgkins(), irp_content_kl_hodgkins_model, and irp_content_h_hodgkins_model that the related models are not reliable for peat.

Bug fixes

• Correct a typo which made it impossible to compute holocellulose_2 with irp_content() (now: irp_predict(), see below).
• irp_content_klh_hodgkins_predict(): The function now checks that spectra have the correct wavenumber range.
• irp_eac_1(), irp_edc_1(): Both functions now also accept ir objects with empty spectra as arguments.

irpeat 0.1.0

• Added a NEWS.md file to track changes to the package.
• Update the documentation (adding missing examples).
• Add pkgdown site.

Bug fixes

• Fix bugs caused by changes in ‘ir’ 0.2.0 (ir::ir_check_ir() no longer is exported, measurement_id is no longer a required column in ir objects).
• Fix bug in irp_content() if variable = "klason_lignin_hodgkins" (no values were computed).

New models

• New models to predict holocellulose (model_holocellulose_2) and Klason lignin (model_klason_lingin_2) contents.

Deprecated functions