private void processLink(String markup, StringBuffer context, HashSet<Integer> bannedTopics) {
// ignore everything that is not in main namespace
if (markup.indexOf(":") > 0) return;
String anchor = markup;
String dest = markup;
int pos = markup.lastIndexOf("|");
if (pos > 0) {
anchor = markup.substring(pos + 1);
dest = markup.substring(0, pos);
}
context.append("\n").append(anchor);
Article art = wikipedia.getArticleByTitle(dest);
if (art != null) {
bannedTopics.add(art.getId());
}
}
/**
* This efficiently identifies sentences within this article that contain links to the given
* target article. The actual text of these sentences can be obtained using {@link
* Page#getSentenceMarkup(int)}
*
* @param art the article of interest.
* @return an array of sentence indexes that contain links to the given article.
*/
public Integer[] getSentenceIndexesMentioning(Article art) {
DbLinkLocationList tmpLinks = env.getDbPageLinkIn().retrieve(art.getId());
if (tmpLinks == null || tmpLinks.getLinkLocations() == null) return new Integer[0];
DbLinkLocation key = new DbLinkLocation(id, null);
int index =
Collections.binarySearch(
tmpLinks.getLinkLocations(),
key,
new Comparator<DbLinkLocation>() {
@Override
public int compare(DbLinkLocation a, DbLinkLocation b) {
return new Integer(a.getLinkId()).compareTo(b.getLinkId());
}
});
if (index < 0) return new Integer[0];
ArrayList<Integer> sentenceIndexes =
tmpLinks.getLinkLocations().get(index).getSentenceIndexes();
return sentenceIndexes.toArray(new Integer[sentenceIndexes.size()]);
}
