wkt polygon validation gives false negatives when spaces follow commas about ala4r HOT 8 CLOSED

Thanks Shawn - actually there's a newish R package from @sckott for handling WKT strings, including a validator - https://github.com/ropensci/wellknown. I'll look at replacing ALA4R's home-rolled dodgy check_wkt with that.

from ala4r.

No worries Ben. That saves me listing another issue regarding ring ordering checks.

Cheers,
Shawn.

from ala4r.

Hmm, the wellknown package is an improvement, but not a total solution. It also thinks that whitespaces are invalid in some cases (e.g. before the comma). I'm not sure if they are valid or not, but it''s kind of a moot point because the ALA servers quite happily accept those WKT strings. So I've also modified the ALA4R functions that take WKT strings, so that WKT strings are NOT checked prior to submitting the request to the ALA. If the request fails (and if the server provides no diagnostic information) then the WKT string is checked and if it fails, a warning is issued that it might have been the problem. The occurrences function always behaved this way, but specieslist and sites_by_species did not.
I think that's about the best we can do for the time being ... let me know if it doesn't seem a suitable solution (or if the ring ordering checks are causing problems).

(ALA4R v1.25 contains these changes)

from ala4r.

Thanks for that. I'll sing out if problems persist.

from ala4r.

@raymondben can you share some example wkt that you refer to that was invalid, and what fxn were you using in wellknown

from ala4r.

@sckott - sure, see ropensci-archive/wellknown#9

from ala4r.

thanks

from ala4r.

ropensci/wellknown now handles the whitespaces, so our check_wkt::is_valid_wkt_polygon also handles them. Further wkt parsing inconsistencies - please notify!

from ala4r.