aboutsummaryrefslogtreecommitdiffstats
path: root/scalding/src/main
diff options
context:
space:
mode:
Diffstat (limited to 'scalding/src/main')
-rw-r--r--scalding/src/main/scala/sandcrawler/Scorable.scala13
1 files changed, 9 insertions, 4 deletions
diff --git a/scalding/src/main/scala/sandcrawler/Scorable.scala b/scalding/src/main/scala/sandcrawler/Scorable.scala
index 736c175..ce4fdca 100644
--- a/scalding/src/main/scala/sandcrawler/Scorable.scala
+++ b/scalding/src/main/scala/sandcrawler/Scorable.scala
@@ -45,12 +45,17 @@ object Scorable {
}
def titleToSlug(title : String) : String = {
- val slug = StringUtilities.removePunctuation(
- StringUtilities.removeAccents(title).split(":")(0).toLowerCase())
- if (slug.isEmpty) {
+ if (title == null || title.isEmpty) {
NoSlug
} else {
- slug
+ val unaccented = StringUtilities.removeAccents(title)
+ // Remove punctuation after splitting on colon.
+ val slug = StringUtilities.removePunctuation((unaccented.split(":")(0).toLowerCase()))
+ if (slug.isEmpty || slug == null) {
+ NoSlug
+ } else {
+ slug
+ }
}
}