lee/heritrix3
1
0
Fork 0
mirror of https://github.com/internetarchive/heritrix3.git synced 2026-09-23 14:15:47 +00:00
Code Issues Packages Projects Releases Wiki Activity

* JerichoExtractorHTMLTest.java

Browse Source 
avoid redundancy by using extractor built in setUp()
    makeExtractor() - call extractor.setExtractorJS(new ExtractorJS()) since we got rid of the static ExtractorJS stuff
    testConditionalComment1() - override to skip the test since it fails with JerichoExtractorHTML
This commit is contained in:
Noah Levitt
2012-12-11 17:00:23 -08:00
parent 85118c59c4
commit 5586d5452d
1 changed files with 20 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/src/test/java/org/archive/modules/extractor/JerichoExtractorHTMLTest.java
+20 -10
View File
@@ -45,6 +45,7 @@ public class JerichoExtractorHTMLTest extends ExtractorHTMLTest {
CrawlMetadata metadata = new CrawlMetadata();
metadata.afterPropertiesSet();
result.setMetadata(metadata);
result.setExtractorJS(new ExtractorJS());
result.afterPropertiesSet();
return result;
}
@@ -68,8 +69,7 @@ public class JerichoExtractorHTMLTest extends ExtractorHTMLTest {
" </select>" +
" <input type=\"submit\" name=\"test\" value=\"Go\">" +
"</form>";
JerichoExtractorHTML ex = (JerichoExtractorHTML)makeExtractor();
ex.extract(curi, cs);
getExtractor().extract(curi, cs);
curi.getOutLinks();
assertTrue(CollectionUtils.exists(curi.getOutLinks(), new Predicate() {
public boolean evaluate(Object object) {
@@ -97,8 +97,7 @@ public class JerichoExtractorHTMLTest extends ExtractorHTMLTest {
" </select>" +
" <input type=\"submit\" name=\"test\" value=\"Go\">" +
"</form>";
JerichoExtractorHTML ex = (JerichoExtractorHTML)makeExtractor();
ex.extract(curi, cs);
getExtractor().extract(curi, cs);
curi.getOutLinks();
assertTrue(!CollectionUtils.exists(curi.getOutLinks(), new Predicate() {
public boolean evaluate(Object object) {
@@ -125,10 +124,9 @@ public class JerichoExtractorHTMLTest extends ExtractorHTMLTest {
" <option value=\"nonselectedOption\">option2</option>" +
" </select>" +
" <input type=\"submit\" name=\"test\" value=\"Go\">" +
"</form>";
JerichoExtractorHTML ex = (JerichoExtractorHTML)makeExtractor();
ex.setExtractOnlyFormGets(false);
ex.extract(curi, cs);
"</form>";
getExtractor().setExtractOnlyFormGets(false);
getExtractor().extract(curi, cs);
curi.getOutLinks();
assertTrue(CollectionUtils.exists(curi.getOutLinks(), new Predicate() {
public boolean evaluate(Object object) {
@@ -142,8 +140,20 @@ public class JerichoExtractorHTMLTest extends ExtractorHTMLTest {
UURI uuri = UURIFactory.getInstance("http://www.example.org");
CrawlURI curi = new CrawlURI(uuri);
CharSequence cs = "<a src=\"http://www.example.com/\" href=\"http://www.archive.org/\"> ";
JerichoExtractorHTML ex = (JerichoExtractorHTML)makeExtractor();
ex.extract(curi, cs);
getExtractor().extract(curi, cs);
assertTrue("not all links found", curi.getOutLinks().size() == 2);
}
/*
* Override of ExtractorHTMLTest method because the test fails with
* JerichoExtractorHTML
*/
@Override
public void testConditionalComment1() throws URIException {
}
@Override
protected JerichoExtractorHTML getExtractor() {
return (JerichoExtractorHTML) extractor;
}
}