stics_wrapper : improve stics outputs management through var.mod completion and variables names checks

[plecharpent]

See if it would be usefull to generate the var.mod file from the wrapper using the variables
name vector passed through the var input argument.

The original var.mod existing in the workspace will be restored at the end of simulations.

[sbuis]

Actually, stics_wrapper generates the var.mod file and re-run STICS in case the required variables are not included in the STICS output files.
The original var.mod file is however not restored. Not sure it is a problem. Let think about an iterative simulation case where a variable not included in the var.mod file is required: if the original var.mod file is restored then the wrapper will systematically run STICS two times.

2 issues however that would be nice to solve:

• some warning messages supposed to inform the user about the generation of a new var.mod file are not displayed (due to foreach loop?)
• the existence of the STICS variables is not checked so that an unexisting variable can be listed in the var.mod file, and in this case, STICS return (null) values for this variable in its output files, so that stics_wrapper believe that this variable actually exists. The existence of the output variables could be checked from the outputs.csv file in stics_wrapper after the first run of the model in case the variable is not found in its output files.

[plecharpent]

SticsRFiles::is_stics_var could be used to check var names, aleady used in gen_varmod but messages may not be displayed as you mentionned. To be improved.