@Override
protected ArrayList<HashMap<String, String>> doInBackground(Void... params) {
ArrayList<HashMap<String, String>> authors = new ArrayList<HashMap<String, String>>();
try {
char l = 'a';
while (l <= 'a') {
URL url = new URL("http://www.liberliber.it/audiolibri/" + l + "/index.htm");
Document doc = Jsoup.parse(url, 5000);
Element e = doc.getElementById("riga02_colonna02");
e = e.getElementsByClass("contenuto_cornice").first();
e = e.getElementsByTag("tbody").first();
e = e.getElementsByTag("tr").get(1);
e = e.getElementsByTag("td").get(1);
e = e.getElementsByTag("ul").first();
for (Element curr : e.getElementsByTag("li")) {
HashMap<String, String> m = new HashMap<String, String>();
Element el = curr.getAllElements().first();
m.put("author", el.text());
m.put("url", el.unwrap().absUrl("href"));
authors.add(m);
}
l++;
}
} catch (Exception e) {
e.printStackTrace();
}
return authors;
}
private void unwrap(final Element element) {
final Set<Element> elementsToUnwrap = new LinkedHashSet<>();
element.traverse(
new NodeVisitor() {
@Override
public void head(Node node, int depth) {
if (node instanceof Element) {
final Element element = (Element) node;
if (element.isBlock()) {
final Set<String> classes = element.classNames();
removeEmpty(classes);
if (classes.isEmpty()) {
elementsToUnwrap.add(element);
}
}
}
}
@Override
public void tail(Node node, int depth) {}
});
for (final Element unwrap : elementsToUnwrap) {
unwrap.unwrap();
}
}
public static String getHighlightedText_math(
String text, String color, String apiElement) // for math after merging with recodoc
{
String highlightBeginning = "<SPAN style=\"BACKGROUND-COLOR: " + color + "\">";
String highlightEnding = "</SPAN>";
Document doc = Jsoup.parse(text);
Elements apiElements = doc.select("clt[api=" + apiElement + "]");
for (Element apielement : apiElements) {
Document tmp = new Document("");
String[] apis = apielement.text().split("\\.");
if (apis.length == 2)
apielement.html(highlightBeginning + apis[0] + highlightEnding + "." + apis[1]);
else apielement.wrap(highlightBeginning);
}
// highlight code snippet
Elements codesnippets = doc.getElementsByTag("pre");
for (Element codesnippet : codesnippets) {
String html = codesnippet.html();
Pattern apielementPattern = Pattern.compile("(?<=\\W)" + apiElement + "(?=\\W)");
Matcher matcher = apielementPattern.matcher(html);
codesnippet.html(matcher.replaceAll(highlightBeginning + apiElement + highlightEnding));
}
// remove clt tags for display
Elements clts = doc.getElementsByTag("clt");
for (Element clt : clts) {
clt.unwrap();
// clt.replaceWith(new TextNode(clt.text(), ""));
}
return doc.html();
}