# itsadug 1.0

## Major changes:

### a. New functions:

- inspect_random: plot random smooths, calculate averages, retrieve values

- plot_topo: plotting EEG too plots, based on fvisgam, pvisgam and plot_diff2

- Function infoMessages: infoMessages('off') to silent all info
messages in the current session (infoMessages('on') will show these
again). Useful when generating knitr reports. Note: the argument
print.summary=FALSE can still be used.

- Function color_contour: for colored contour plot.

- Function plotSurface: for plotting a surface plot from a data frame
with fitted values based on the predict functions get_modelterm,
get_predictions.

### b. Bugs fixed:
- compareML with GCV or other (not REML, fREML, or ML) gives correct
EDF and different output

- Bug in get_difference, which caused errors in plot_diff and
plot_diff2

- Bug from acf_resid removed 

- resid_gam now also works for matrix input (cbind predictors).

- Fixed bug in check_resid: " Error in plot.window(...) : need finite
'ylim' values "

- gamtabs now also works for parametric models (thanks to Christian and
Samuel)

- fadeRug reimplemented, bugs fixed, examples added.

- Fixed bug in addInterval: color vectors went wrong with zero length
elements, now fixed.


### c. New functionality / arguments of existing functions

- plot_smooths: argument plot_all added to plot all levels of a certain
predictor

- Added optional argument transform in plot_smooth and fvsigam to
transform model values.

- plot_diff is completely reimplemented to add some functionality. Now
it’s possible to specify more than one predictor to define groups; The
argument cond is added for specifying other predictors as a named list.
The argument rm.ranef is added to cancel random effects. Random slopes
and random smooths may change the difference curve. 

- Reimplemented plot_diff2: It's now possible to remove random effects, and to
change plot settings.

- get_difference: changed input arguments cond1 and cond2 into one
argument comp, to keep the input arguments more consistent between
different functions. In addition, it’s now possible to use rm.ranef,
and to specify more than one predictor to define groups.

- Labels added to plots that are produced by fvisgam, pvisgam,
plot_diff, plot_diff2, and plot_smooth to make a clear distinction
between the plots. The argument hide.label could be used to remove
these labels.

- Function get_random changed: Now it only outputs the coefficients of random intercepts and slopes. Use the new function inspect_random for extracting random factor smooths.


## Minor changes:

### a. Vignettes:

- Changed vignettes, because vignette “plot functions” became to big.
Will be published on my website soon. Instead, vignette(“overview”) and
vignette(“acf”) for short descriptions.
n)

### b. Other changes:

- acf_resid, acf_plot, and acf_n_plots now can plot confidence lines.
The argument return_all in acf_resid and acf_plot provides more
advanced output. See vignette("acf") for all possibilities that these
ACF functions offer.

- summary output is changed: "Canceled as random effect." was confusing
in some cases. Now changed in "Might be canceled as random effect,
check below."

- resid_gam also possible to use for models without AR1 included.

- resid_gam and acf_resid should now also work for lme4 linear mixed
models.

- find_difference now optionally returns a vector with logical values
instead of list with start and end points (use as.vector=TRUE)

- Added possibility to extend or change xlim and ylim for pvisgam,
fvisgam, and plot_diff2


