Merge 27d9703 into e29bced

intenthq · Sep 15, 2015 · 09e7de2 · 09e7de2
2 parents e29bced + 27d9703
commit 09e7de2
Show file tree

Hide file tree

Showing 2 changed files with 43 additions and 1 deletion.
diff --git a/README.md b/README.md
@@ -1,4 +1,4 @@
-#Gander [![Build Status](https://img.shields.io/travis/intenthq/gander.svg)](https://travis-ci.org/intenthq/gander) [![Coverage Status] (https://img.shields.io/coveralls/intenthq/gander.svg)](https://coveralls.io/github/intenthq/gander?branch=master) [![Maven Central](https://img.shields.io/maven-central/v/com.intenthq/gander_2.11.svg)](http://search.maven.org/#search%7Cga%7C1%7Cg%3A%22com.intenthq%22%20AND%20a%3A%22gander_2.11%22) [![Join the chat at https://gitter.im/intenthq/gander](https://img.shields.io/badge/gitter-join%20chat-green.svg)](https://gitter.im/intenthq/gander)
+#Gander [![Build Status](https://img.shields.io/travis/intenthq/gander/master.svg)](https://travis-ci.org/intenthq/gander) [![Coverage Status] (https://img.shields.io/coveralls/intenthq/gander.svg)](https://coveralls.io/github/intenthq/gander?branch=master) [![Maven Central](https://img.shields.io/maven-central/v/com.intenthq/gander_2.11.svg)](http://search.maven.org/#search%7Cga%7C1%7Cg%3A%22com.intenthq%22%20AND%20a%3A%22gander_2.11%22) [![Join the chat at https://gitter.im/intenthq/gander](https://img.shields.io/badge/gitter-join%20chat-green.svg)](https://gitter.im/intenthq/gander)
 
 **Gander is a scala library that extracts metadata and content from web pages.**
 

diff --git a/src/test/scala/com/intenthq/gander/ContentExtractorSpec.scala b/src/test/scala/com/intenthq/gander/ContentExtractorSpec.scala
@@ -17,6 +17,48 @@ class ContentExtractorSpec extends Specification {
     }
   }
 
+  "extractCanonicalLink" >> {
+    "should return none if no link found" >> {
+      val html =
+        """<html lang="ca">
+          |  <head>
+          |  </head>
+          |<body></body></html>""".stripMargin
+      extractCanonicalLink(Jsoup.parse(html)) must beNone
+    }
+
+    "should extract the canonical link from the meta tag" >> {
+      val html =
+        """<html lang="ca">
+          |  <head>
+          |    <link rel="canonical" href="http://example.com/canonical">
+          |    <meta property="og:url" content="http://example.com/og" />
+          |    <meta name="twitter:url" content="http://example.com/twitter" />
+          |  </head>
+          |<body></body></html>""".stripMargin
+      extractCanonicalLink(Jsoup.parse(html)) must beSome("http://example.com/canonical")
+    }
+    "should extract the facebook og:url meta tag" >> {
+      val html =
+        """<html lang="ca">
+          |  <head>
+          |    <meta property="og:url" content="http://example.com/og" />
+          |    <meta name="twitter:url" content="http://example.com/twitter" />
+          |  </head>
+          |<body></body></html>""".stripMargin
+      extractCanonicalLink(Jsoup.parse(html)) must beSome("http://example.com/og")
+    }
+    "should extract the twitter:url meta tag" >> {
+      val html =
+        """<html lang="ca">
+          |  <head>
+          |    <meta name="twitter:url" content="http://example.com/twitter" />
+          |  </head>
+          |<body></body></html>""".stripMargin
+      extractCanonicalLink(Jsoup.parse(html)) must beSome("http://example.com/twitter")
+    }
+  }
+
   "extractLang" >> {
     "should extract lang from html tag and give priority to it" >> {
       val html =