add na_to_zero function
add rmoutliers, fct_to_filter, rmlevels,boolean_filter,onlyfiteddata,del_boolean arguments to lm_table and lm_resid
add m3ha to sprs,ss_diffs,strs;
add rqm filter and option to show filtered data only, or original and filtered to lm_table;
replace order function call with dplyr’s arrange, due to order’s behavior in certain situations;
add shiny and miniUI dependency;
add lm_resid and lm_resid_group experimental functions. These output shiny gadgets;
changes in sprs,ss_dffs,classify_site,dom_height,fit_clutter,plot_summarise,smalianwb,smalianwob,tree_summarise due to dplyr updates (due to .add argument);
updated nls_table and lm_table to work with tibble 3.0.0;
new function: rm_empty_cols; this function is used internally to better handle optional variables;
added a new dependency, forcats;
new function: ident_model_full; this function runs the complete identity model test;
fixed a bug where lm_table and nls_table would sometimes change the order of coefficients;
graybill_f now accepts vectors as input;
average_tree_curve now has a color argument, that allows variables to be mapped as colors;
average_tree_curve now plots a mirrored plot, as to resemble a tree (mirror argument) and has an argument to remove the equation from the plot;
rmse_perc and bias_perc now have a na.rm argument (which is TRUE by default);
updated tree_summarise to sum volume with and without bark using 2 new arguments;
fixed vol_summarise handling of optional variables;
new function: class_center;
diameter_class and bdq_meyer can now handle all values of class intervals correctly.
Graybill_F is now graybill_f;
removed unnecessary imports;
updated description;
converted all example data to tibbles;
removed head command from examples, since it’s not necessary anymore;
fixed some typos;
added a gray_scale argument to est_clutter function;
fixed a bug where if a dataframe contained NA, resid_plot would fail;
vol_summarise now outputs ungroup data;
fixed vol_summarise name for average form factor value;
fixed the example of lm_table;
fixed the example of strs; number of plots needed to be sampled was wrong;
changed exfm19 to a data by tree;
added 6 vignettes in Portuguese;
added 6 vignettes in English.
tree_summarise is now improved, and keeps all variables from data;
strs can now handle special characters;
added a new example data, with revenue values for a eucalyptus forest;
added a new function, npv_irr, for net present value, irr calculation and sensibility evaluation;
fixed a bug in huberwob, which made a wrong call to df;
fixed bugs in various functions, with a call for is.na with a 2 length vector.