public static void newUrl(String url) throws Exception {
String body = Robot.get(url);
Page p = new Page();
p.setTmsp("2011-03-11 11:00:00");
p.setUrl(url);
p.setBody(body);
p.insert();
}
public static Page next(ResultSet rs) throws Exception {
Page p = null;
if (rs.next()) {
p = new Page();
p.setTmsp(rs.getString("tmsp"));
p.setUrl(rs.getString("url"));
p.setBody(rs.getString("body"));
}
return p;
}
// Chamando Robot
public static void robot() throws Exception {
Database db = new Database();
db.connect();
ResultSet rs = Page.findAll(db);
Page p = null;
while ((p = Page.next(rs)) != null) {
String body = Robot.get(p.getUrl());
// procurar por urls dentro do body
// buscar por essas paginas
// String expr = "href=\"([^\"]*)";
String ereg = "href=\"https{0,1}:\\/\\/([^\"]*)\"";
Pattern pt = Pattern.compile(ereg);
Matcher m = pt.matcher(body);
while (m.find()) {
System.out.println(m.group());
String[] _url = m.group().split("\"");
Page.newUrl(_url[1]);
}
p.setBody(body);
p.update();
}
db.close();
}
public static void test02() throws Exception {
Page.robot();
}
public static void test01() throws Exception {
Page p = new Page();
p.setTmsp("2011-03-04 10:30:00");
p.setUrl("http://www.google.com");
p.setBody("my body");
p.insert();
Page e = new Page("2011-03-04 10:31:00", "http://www.usp.br", "o corpo");
e.insert();
e.setBody("usp usp usp");
e.update();
Database db = new Database();
db.connect();
ResultSet rs = Page.findAll(db);
/*
while (rs.next()) {
System.out.println(rs.getString("tmsp")
+", "+ rs.getString("url")
+", "+ rs.getString("body"));
}*/
Page p0 = null;
while ((p0 = Page.next(rs)) != null) {
System.out.println(p0.getTmsp());
System.out.println(p0.getUrl());
System.out.println(p0.getBody());
}
db.close();
p.remove();
e.remove();
}
