quick fix for issue where '\s+' is passed as mfile_sep

[briochh]

Flopy written external files are often written with a leading space in the first column. If these files have a .csv extension, adding to pars to PstFrom would require mfile_sep=r"\s+" to support consecutive delimiters. This was causing issues downstream when writing models at apply time (r"\s+" is not an accepted separator when writing!). For now scrubbing "\s+" to store ' ' in apply info file.

Broader question -- should we be assuming mfile_sep=" " means mfile_sep=r"\s+" when passed (as we do if the model file extension is not .csv) or are we better honouring what the user explicitly passes here (if they are passing the keyword arg)? My preference is for the latter (as that is what we are currently doing, it provides more flexibility, explicit > implicit, and now mult white space can now be appropriately handled.)

[coveralls]

coverage: 79.089% (+0.06%) from 79.026%
when pulling bd1b391 on briochh:hotfix_parsep
into 41e0f85 on pypest:develop.