phylopic_uid & phylopic_uid_item return "subscript out of bounds" errors when a species is not found

[docmanny]

When using phylopic_uid, sometimes it will fail with error Error in jsonlite::fromJSON(url)$result[[1]] : subscript out of bounds whenever Phylopic doesn't have the species in their database. For a small set of species names, troubleshooting this isn't too hard, but for longer species lists, figuring out which species are causing the problem can add up to a significant timesink.

Specifically, the problem lies in the ggimage internal function phylopic_uid_item, and its assumption that you will always get at least one item from jsonlite::fromJSON(url)$result. It would be safer to first save the object to a variable, and then check for the length before proceeding.

My proposed fix would be to add an if-else statement checking the length of the results item as follows:

phylopic_uid_item <- function (name) 
{
    x <- gsub("[^a-zA-Z]+", "+", name)
    url <- paste0("http://phylopic.org/api/a/name/search?text=", 
                  x, "&options=scientific+json")
    results <- jsonlite::fromJSON(url)$result   
    if (length(results)==0){
        return(NA)
    } else{
    res <- results[[1]]
    for (id in res$uid) {
        uid <- phylopic_valid_id(id)
        if (!is.na(uid)) 
            break
    }
    return(uid)}
}

As a minimal test case, you can compare the output of phylopic_uid("Procavia_capensis") (which always works) with phylopic_uid("Chlorocebus_sabaeus") (which fails currently, but returns NA with my solution.

If this works well for you, then I can also submit a pull request with the change.

Edit:
Additionally, this fix would require changing "vapply" to "sapply" in phylopic_uid to allow NA values.

[GuangchuangYu]

PR is welcome :)