public JobData parse(Document doc) {
String keyword = doc.select("meta[name=keywords]").first().attr("content");
String[] keywordArr = keyword.split(",");
String uniqueId = extractUniqueId(doc);
String companyName = keywordArr[0];
String jobTitle = keywordArr[1];
List<String> cityList = extractCities(doc);
Date date = extractDate(doc);
String summary = summary(doc);
JobData jobData = null;
try {
jobData =
new JobData(
trim(SOURCE),
trim(uniqueId),
trim(companyName),
trim(jobTitle),
cityList,
trim(summary),
date,
trim(doc.baseUri()),
"");
} catch (Exception ex) {
System.err.println(doc.baseUri() + " , " + ex.getMessage());
}
return jobData; // To change body of implemented methods use File | Settings | File Templates.
}
public void index(final String title, String wikitext, Document doc) {
try {
final String hash = new StringHash(wikitext).hash();
// parse input with JSOUP
class Walker {
private int secIndex;
Walker(int secIndex) {
this.secIndex = secIndex;
}
public int walk(Element el) {
Elements children = el.children();
String tagName = el.tagName().toLowerCase();
if (tagName.matches("h[1-6]")) {
secIndex++;
String secName = el.text();
String key =
"SectionTitle" + "/" + title + "/" + hash + "/" + new Integer(secIndex).toString();
String value = secNameFilter(secName);
dataStore.put(key, value);
}
for (Element child : children) {
secIndex = (new Walker(secIndex)).walk(child);
}
return secIndex;
}
}
(new Walker(0)).walk(doc.body());
} catch (Exception e) {
System.err.println(
"SectionTitleIndexer.index failed due to exception: '" + e.toString() + "'");
}
}
