public static String extractContent(String url) {
try {
Connection connection = Jsoup.connect(url);
connection.userAgent(USER_AGENT);
connection.followRedirects(true);
connection.timeout(GET_TIMEOUT);
long start = System.currentTimeMillis();
Connection.Response response = connection.execute();
long diff = System.currentTimeMillis() - start;
int responseCode = response.statusCode();
if (response.statusCode() == OK) {
String body = response.body();
Logger.info(
"%s retrieved, content length %d, time %s sec.",
url, body.length(), FormatUtil.millis2Seconds(diff));
return response.body();
} else {
Logger.error("%s returned %d", url, responseCode);
return "";
}
} catch (IOException e) {
Logger.error(e, "%s cannot be read.", url);
return "";
}
}
@Override
protected String doProcess(File htmlfile, String originalUrl, Intent intent) {
try {
// String charset = "utf-8";
Connection coon = HttpConnection.connect(originalUrl);
coon.followRedirects(
false); // we don't want it be redirected to other page,example: 10.254.7.4
Document doc = coon.get();
Element head = doc.head();
Element body = doc.body();
if (body.children().size() == 0) {
Log.e(TAG, "body has no child with url=" + originalUrl);
return PROCESS_FAILED_URL;
}
/*
Elements meta = head.select("meta");
if(!meta.isEmpty()){
Element m = meta.get(0);
String content = m.attr("content");
String attr = content.substring(content.indexOf("charset=")+8);
if(!attr.trim().isEmpty()){
charset = attr;
}
}
*/
Elements base = head.select("base");
if (base.isEmpty()) {
String b = head.baseUri();
Attributes attrs = new Attributes();
attrs.put("href", b);
ArrayList<Element> a = new ArrayList<>();
a.add(new Element(Tag.valueOf("base"), b, attrs));
head.insertChildren(0, a);
}
Element div = doc.select("div.content-main").first();
if (div == null) {
Log.e(TAG, "not found specific element with url=" + originalUrl);
return PROCESS_FAILED_URL;
}
Element title = div.select("h1.title").first();
title.remove();
body.empty();
ArrayList<Element> a = new ArrayList<>();
a.add(div);
body.insertChildren(0, a);
int g = 0;
while (g < 2) { // try two times.
if (FileUtil.saveStringToFile(doc.toString(), htmlfile, false)) {
break;
}
g++;
}
if (g < 2) return StringUtils.file2Url(htmlfile, PROCESS_FAILED_URL);
Log.e(TAG, "save html to file failed with url=" + originalUrl);
} catch (MalformedURLException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
return PROCESS_FAILED_URL;
}
