aboutsummaryrefslogtreecommitdiffstats
path: root/scalding/src/test/scala/sandcrawler/CrossrefScorableTest.scala
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2018-08-23 19:35:23 -0700
committerBryan Newbold <bnewbold@archive.org>2018-08-23 19:36:23 -0700
commit98e67e291132b10a0ca698ad4ff754acc0c22121 (patch)
tree940d436787c14548eacf6b86d3670a9f3ef3b1ad /scalding/src/test/scala/sandcrawler/CrossrefScorableTest.scala
parent715a35715609d8cbacff53dd5c7c1715c53a55f8 (diff)
downloadsandcrawler-98e67e291132b10a0ca698ad4ff754acc0c22121.tar.gz
sandcrawler-98e67e291132b10a0ca698ad4ff754acc0c22121.zip
add a content-type filter for crossref works
Diffstat (limited to 'scalding/src/test/scala/sandcrawler/CrossrefScorableTest.scala')
-rw-r--r--scalding/src/test/scala/sandcrawler/CrossrefScorableTest.scala9
1 files changed, 9 insertions, 0 deletions
diff --git a/scalding/src/test/scala/sandcrawler/CrossrefScorableTest.scala b/scalding/src/test/scala/sandcrawler/CrossrefScorableTest.scala
index 0cb12ee..f598cae 100644
--- a/scalding/src/test/scala/sandcrawler/CrossrefScorableTest.scala
+++ b/scalding/src/test/scala/sandcrawler/CrossrefScorableTest.scala
@@ -74,6 +74,8 @@ class CrossrefScorableTest extends FlatSpec with Matchers {
val CrossrefStringWithoutTitle = CrossrefString.replace("title", "nottitle")
val MalformedCrossrefString = CrossrefString.replace("}", "")
val CrossrefStringWithNoAuthors = CrossrefString.replace("<<TITLE>>", "Some Valid Title").replace("author", "no-author")
+ val CrossrefStringWrongType = CrossrefString.replace("<<TITLE>>", "Some Valid Title").replace("journal-article", "other")
+ val CrossrefStringNoType = CrossrefString.replace("<<TITLE>>", "Some Valid Title").replace("type", "not-type")
// Unit tests
"CrossrefScorable.jsonToMapFeatures()" should "handle invalid JSON" in {
@@ -139,4 +141,11 @@ class CrossrefScorableTest extends FlatSpec with Matchers {
it should "return false for invalid JSON" in {
CrossrefScorable.keepRecord(CrossrefStringWithoutTitle) shouldBe false
}
+
+ it should "handle content types" in {
+ val resultWrong = CrossrefScorable.jsonToMapFeatures(CrossrefStringWrongType)
+ resultWrong.slug shouldBe Scorable.NoSlug
+ val resultMissing = CrossrefScorable.jsonToMapFeatures(CrossrefStringNoType)
+ resultMissing.slug shouldBe Scorable.NoSlug
+ }
}