Predict soil properties from a spectral data frame using soilVAE models

Predict soil properties from a spectral data frame using soilVAE models

Usage

predict_soil(
  df,
  family_id,
  properties = NULL,
  model_dir = getOption("autoSpectra.model_dir", "models"),
  disable_pp = FALSE
)

Arguments

df

Data frame with a Soil_ID column and numeric spectral columns named by wavelength (nm) or wavenumber (cm-1)

family_id

Family identifier: "OSSL_VisNIR" or "OSSL_MIR"

properties

Character vector of OSSL L1 property keys to predict. Default (NULL) predicts all properties available for the family.

model_dir

Root directory where trained models are stored

disable_pp

Logical; skip spectral preprocessing (debug only)

Value

Data frame with Soil_ID and one column per predicted property