more updates to effervescence

 - greediness: preferentially take first class (after comma or semicolon most likely to be matrix eff)
 - allow ranges of classes separated by "to" (like drainage class)
 - note there are a wide range of narrative comments still not yet handled e.g. "slightly effervescent but strongly effervescent in spots"

R/parseOSD_functions.R

@@ -207,8 +207,8 @@
 
 # vectorized parsing of effervescence class
 .parse_eff_class <- function(x) {
-  SoilKnowledgeBase:::.zerochar_to_na(gsub("^.*(very [a-z]+ effervescen[tce]+).*$|^.*\\b([a-z]+ ?effervescen[tce]+).*$|.*",
-                                           "\\1\\2", x, ignore.case = TRUE))
+  SoilKnowledgeBase:::.zerochar_to_na(gsub("^.*[;,]? \\b([a-z]+ ?effervescen[tce]+ to [a-z]+ ?effervescen[tce]+).*$|^.*[;,] \\b(very [a-z]+ effervescen[tce]+).*$|^.*[;,] \\b([a-z]+ ?effervescen[tce]+).*$|^.*[;,]? \\b(very [a-z]+ effervescen[tce]+).*$|^.*[;,]? \\b([a-z]+ ?effervescen[tce]+).*$|.*",
+                                           "\\1\\2\\3\\4\\5", x, ignore.case = TRUE))
   # factors cannot be preserved in JSON output, and wont work for multiple classes/ranges of classes
 }