So, What Are You? ..A Plant? ..An Animal? -- Nope, I'm a Fungus!

Lately I had a list of about 1000 species names and I wanted to filter out only the plants as that is where I come from. I knew that Scott Chamberlain has put together the ritis package which obviously can do such things. However, I knew of ITIS before and was keen to give it a shot..

Here's what I've come up with (using the ITIS API, updated on 11. Dec 2012, previous version had a flaw with indefinite matches.. Should be ok now. However, there are of course species that are not covered by the database, i.e. Ixodes, see below):

library(XML)
get_tsn <- function(sp_name) {
           require(XML)
           units <- tolower(unlist(strsplit(sp_name, " ")))

           # valid string?
           if (length(units) > 2) { stop("...No valid search string submitted (two words seperated by one space)!") }

           itis_xml <- htmlParse(paste("http://www.itis.gov/ITISWebService/services/ITISService/searchByScientificName?srchKey=", 
                                       sp_name, sep=""))
           tsn <- xpathSApply(itis_xml, "//tsn", xmlValue)
           unitname1 <- tolower(gsub("\\s+", "", xpathSApply(itis_xml, "//unitname1", xmlValue)))
           unitname2 <- tolower(gsub("\\s+", "", xpathSApply(itis_xml, "//unitname2", xmlValue)))
           unitname3 <- tolower(gsub("\\s+", "", xpathSApply(itis_xml, "//unitname3", xmlValue)))

           # sp_name = only Genus, get tsn were sp_name matches perfectly and unitname2 (lower level taxon) is absent 
           if (length(units) == 1) {
               return(tsn[tolower(sub("\\s+", "", unitname1)) == tolower(sp_name) & unitname2 == ""]) }

           # sp_name = Genus and Epitheton, get tsn where both match perfectly and unitname3 (lower level taxon) is absent 
           if (length(units) == 2) {
               return(tsn[unitname1 == units[1] & 
                          unitname2 == units[2] &
                          nchar(unitname3) == 0]) }
           }

get_kngdm <- function(tsn) {
                   kngdm <- xpathSApply(htmlParse(paste("http://www.itis.gov/ITISWebService/services/ITISService/getKingdomNameFromTSN?tsn=", 
                                                       tsn, sep="")), 
                                                  "//kingdomname", xmlValue)
           return(kngdm)
           }

get_tsn_kngdm <- function(x) {y = get_tsn(x)
                              z = get_kngdm(y)
                              return(list(Name = x, TSN = y, Kingdom = z))
                              }

# I had some API-related errors (I guess it was mysteriously not answering in 
# some cases). I couldn't resolve this and thus implemented tryCatch
get_tsn_kngdm_try <- function(x) tryCatch(get_tsn_kngdm(x), error = function(e) NULL)

sp_names <- c("Clostridium", "Physcia", "Ixodes", "LYNX", "Homo sapiens", "Canis lupus")

system.time(result <- data.frame(do.call(rbind, lapply(sp_names, FUN = get_tsn_kngdm_try))))
result

system.time(result <- data.frame(do.call(rbind, lapply(sp_names, FUN = get_tsn_kngdm_try))))
#
# result
#        User      System verstrichen 
#        1.54        0.01       33.66 
#           Name    TSN  Kingdom
# 1  Clostridium 555645   Monera
# 2      Physcia  14024    Fungi
# 3        Viola  22030  Plantae
# 4       Ixodes                
# 5         LYNX 180581 Animalia
# 6 Homo sapiens 180092 Animalia
# 7  Canis lupus 180596 Animalia
#