NEWS

flexFitR 1.2.5

Added a new vignette, Model Selection.
A model_selection function was created to automate the model selection process.
New default option for performance. "AICc" was set as default.

Added a new vignette, Modeling soybean canopy cover.
Added the example dataset dt_soybean_22, containing soybean canopy-cover time-series data from Keller et al. (2026).
The print() method for modeler objects now reports the parameter box constraints used during model fitting.

plot.modeler() no longer produces the "Ignoring unknown labels" message with ggplot2 >= 4.0 for plot types 4, 5, and 6. Legend labels are now added only when the corresponding aesthetics are mapped.

compute_tangent() function added to compute tangent line(s) for a modeler object.
inverse_predict.modeler() S3 method added to calculate inverse predictions for modeler objects.
update.modeler() S3 method added to refit a model of class modeler.
Adding fn_lin_logis(), fn_quad_plat() and fn_quad_pl_sm().
predict.modeler() includes parallel and workers to allow for parallel computing.

When evaluating several methods in modeler(), Jacobian and Hessian are computed only for the best method.
Now functions are required to be vectorized (faster execution).
Renaming fn_lin_plat() function.
The modeler() function now uses optimr instead of opm for faster execution.
plot.modeler() includes linewidth argument to increase size in geom lines.

Removed methods that required hessian matrix (snewton, snewtonm, snewtm) in list_methods().
Fixed issue when combining fitted values in modeler().

fitted.modeler() S3 method added to extract fitted values from modeler objects.
residuals.modeler() S3 method added to extract residuals from modeler objects.
augment() function added to calculate influence measures (Cook's distance, leverage values, standardized residuals, studentized residuals).
c.modeler() S3 method added to combine modeler objects.
subset.modeler() S3 method added to subset modeler objects.
performance() function added to evaluate the performance of several models.
plot.performance() S3 method to plot an object of class performance.

modeler() adds the function name (fn_name) in every output table.
modeler() no longer returns function call.
plot.modeler() includes add_ribbon_pi and add_ribbon_ci arguments for prediction and confidence intervals.
metrics() returns R2 instead of r_squared.

Fixed conflict of modeler() with upcoming version of future.
Fixed increase dependency to R (>=4.1).
Fixed regression function not found in the environment when running in parallel.