Example usage for org.jsoup.nodes Document body

List of usage examples for org.jsoup.nodes Document body

Introduction

In this page you can find the example usage for org.jsoup.nodes Document body.

Prototype

public Element body() 

Source Link

Document

Accessor to the document's body element.

Usage

From source file:edu.ucla.cs.scai.swim.qa.ontology.dbpedia.tipicality.DbpediaCsvDownload.java

public static void main(String args[]) throws FileNotFoundException, IOException {
    Document doc = null;
    try {// w  ww . ja v a  2s . c o  m
        doc = Jsoup.connect(DBpediaOntology.DBPEDIA_CLASSES_URL).get();
    } catch (IOException ex) {
        Logger.getLogger(DBpediaOntology.class.getName()).log(Level.SEVERE, null, ex);
    }

    download(doc.body().children().get(1).children().get(1));
}

From source file:akori.AKORI.java

public static void main(String[] args) throws IOException, InterruptedException {
    System.out.println("esto es AKORI");

    URL = "http://www.mbauchile.cl";
    PATH = "E:\\NetBeansProjects\\AKORI\\";
    NAME = "mbauchile.png";
    // Extrar DOM tree

    Document doc = Jsoup.connect(URL).timeout(0).get();

    // The Firefox driver supports javascript 
    WebDriver driver = new FirefoxDriver();
    driver.manage().window().maximize();
    System.out.println(driver.manage().window().getSize().toString());
    System.out.println(driver.manage().window().getPosition().toString());
    int xmax = driver.manage().window().getSize().width;
    int ymax = driver.manage().window().getSize().height;

    // Go to the URL page
    driver.get(URL);// w  w  w.ja  v a 2 s .  com

    File screen = ((TakesScreenshot) driver).getScreenshotAs(OutputType.FILE);
    FileUtils.copyFile(screen, new File(PATH + NAME));

    BufferedImage img = ImageIO.read(new File(PATH + NAME));
    //Graphics2D graph = img.createGraphics();

    BufferedImage img1 = new BufferedImage(xmax, ymax, BufferedImage.TYPE_INT_ARGB);
    Graphics2D graph1 = img.createGraphics();
    double[][] matrix = new double[ymax][xmax];
    BufferedReader in = new BufferedReader(new FileReader("et.txt"));
    String linea;
    double max = 0;
    graph1.drawImage(img, 0, 0, null);
    HashMap<String, Integer> lista = new HashMap<String, Integer>();
    int count = 0;
    for (int i = 0; (linea = in.readLine()) != null && i < 10000; ++i) {
        String[] datos = linea.split(",");
        int x = (int) Double.parseDouble(datos[0]);
        int y = (int) Double.parseDouble(datos[2]);
        long time = Double.valueOf(datos[4]).longValue();
        if (x >= xmax || y >= ymax)
            continue;
        if (time < 691215)
            continue;
        if (time > 705648)
            break;
        if (lista.containsKey(x + "," + y))
            lista.put(x + "," + y, lista.get(x + "," + y) + 1);
        else
            lista.put(x + "," + y, 1);
        ++count;
    }
    System.out.println(count);
    in.close();
    Iterator iter = lista.entrySet().iterator();
    Map.Entry e;
    for (String key : lista.keySet()) {
        Integer i = lista.get(key);
        if (max < i)
            max = i;
    }
    System.out.println(max);
    max = 0;
    while (iter.hasNext()) {
        e = (Map.Entry) iter.next();
        String xy = (String) e.getKey();
        String[] datos = xy.split(",");
        int x = Integer.parseInt(datos[0]);
        int y = Integer.parseInt(datos[1]);
        matrix[y][x] += (int) e.getValue();
        double aux;
        if ((aux = normalMatrix(matrix, y, x, ((int) e.getValue()) * 4)) > max) {
            max = aux;
        }
        //normalMatrix(matrix,x,y,20);
        if (matrix[y][x] > max)
            max = matrix[y][x];
    }
    int A, R, G, B, n;
    for (int i = 0; i < xmax; ++i) {
        for (int j = 0; j < ymax; ++j) {
            if (matrix[j][i] != 0) {
                n = (int) Math.round(matrix[j][i] * 100 / max);
                R = Math.round((255 * n) / 100);
                G = Math.round((255 * (100 - n)) / 100);
                B = 0;
                A = Math.round((255 * n) / 100);
                ;
                if (R > 255)
                    R = 255;
                if (R < 0)
                    R = 0;
                if (G > 255)
                    G = 255;
                if (G < 0)
                    G = 0;
                if (R < 50)
                    A = 0;
                graph1.setColor(new Color(R, G, B, A));
                graph1.fillOval(i, j, 1, 1);
            }
        }
    }
    //graph1.dispose();

    ImageIO.write(img, "png", new File("example.png"));
    System.out.println(max);

    graph1.setColor(Color.RED);
    // Extraer elementos
    Elements e1 = doc.body().getAllElements();
    int i = 1;
    ArrayList<String> tags = new ArrayList<String>();
    for (Element temp : e1) {

        if (tags.indexOf(temp.tagName()) == -1) {
            tags.add(temp.tagName());

            List<WebElement> query = driver.findElements(By.tagName(temp.tagName()));
            for (WebElement temp1 : query) {
                Point po = temp1.getLocation();
                Dimension d = temp1.getSize();
                if (d.width <= 0 || d.height <= 0 || po.x < 0 || po.y < 0)
                    continue;
                System.out.println(i + " " + temp.nodeName());
                System.out.println("  x: " + po.x + " y: " + po.y);
                System.out.println("  width: " + d.width + " height: " + d.height);
                graph1.draw(new Rectangle(po.x, po.y, d.width, d.height));
                ++i;
            }
        }
    }

    graph1.dispose();
    ImageIO.write(img, "png", new File(PATH + NAME));

    driver.quit();

}

From source file:dk.dma.msiproxy.common.util.TextUtils.java

/**
 * Converts the text from html to plain text
 * @param html the html//  ww  w  . j  av a  2  s  .  c o  m
 * @return the plain text version
 */
public static String html2txt(String html) {
    try {
        Document doc = Jsoup.parse(html);
        return new HtmlToPlainText().getPlainText(doc.body());
    } catch (Exception e) {
        // If any error occurs, return the original html
        return html;
    }
}

From source file:com.company.ComponentParser.java

public static Compo Parse(String url) throws Exception {
    InputStream inputStream = new URL(url).openStream();
    String content = IOUtils.toString(inputStream);

    Document document = Jsoup.parse(content);
    Element body = document.body();
    Elements elements = body.select(".grid");
    Compo compo = new Compo();

    Elements tds = elements.get(1).select("th"); //System.out.println(content);

    tds.forEach(element -> {/*www  . j  a v  a  2  s .c  om*/

        //  System.out.println(element.text());
        switch (element.text()) {
        case "License":
            compo.setLicense(element.nextElementSibling().text());
            break;
        case "Categories":
            compo.setCategories(element.nextElementSibling().text());
            break;
        case "HomePage":
            compo.setHomePage(element.nextElementSibling().select("a").text());
            break;
        case "Date":
            compo.setDate(element.nextElementSibling().text());
            break;
        case "Repository":
            compo.setRepository(element.nextElementSibling().text());
            break;
        case "Usages":
            compo.setUsage(element.nextElementSibling().text());
            break;
        }
    });

    return compo;
}

From source file:net.noday.core.dnspod.Dnspod.java

public static String getApiVersion() throws IOException {
    Document doc = Jsoup.connect(url_version).data(data).userAgent(user_agent).post();
    return doc.body().text();
}

From source file:com.mycollab.core.utils.StringUtils.java

/**
 * @param value// ww  w .j a v  a2 s  .  c  o  m
 * @return
 */
public static String formatRichText(String value) {
    if (isBlank(value)) {
        return "";
    }

    value = Jsoup.clean(value,
            relaxed().addTags("img")
                    .addAttributes("img", "align", "alt", "height", "src", "title", "width", "style")
                    .addProtocols("img", "src", "http", "https"));
    Document doc = Jsoup.parse(value);
    Element body = doc.body();
    replaceHtml(body);
    String html = body.html();
    return html.replace("\n", "");
}

From source file:com.mycollab.core.utils.StringUtils.java

private static void replaceHtml(Node element) {
    List<Node> elements = element.childNodes();
    Pattern compile = Pattern.compile("(?:https?|ftps?)://[\\w/%.-][/\\??\\w=?\\w?/%.-]?[/\\?&\\w=?\\w?/%.-]*");
    for (int i = elements.size() - 1; i >= 0; i--) {
        Node node = elements.get(i);
        if (node instanceof TextNode) {
            String value = ((TextNode) node).text();
            Matcher matcher = compile.matcher(value);
            if (matcher.find()) {
                value = value.replaceAll(
                        "(?:https?|ftps?)://[\\w/%.-][/\\??\\w=?\\w?/%.-]?[/\\?&\\w=?\\w?/%.-]*",
                        "<a href=\"$0\" target=\"_blank\">$0</a>");
                Document newDoc = Jsoup.parse(value);
                List<Node> childs = newDoc.body().childNodes();
                for (int j = 0; j < childs.size(); j++) {
                    Node childNode = childs.get(j).clone();
                    node.before(childNode);
                }/*from w  w w. j a v a 2 s .c  o  m*/
                node.remove();
            }
        }
    }
}

From source file:net.noday.core.dnspod.Dnspod.java

public static String domainCreate(Domain obj) {
    Document doc;
    try {//from w w  w  . j  av  a2s .c om
        doc = Jsoup.connect(url_domainCreate).data(data).data("domain", obj.getName()).userAgent(user_agent)
                .post();
        JSONObject o = JSON.parseObject(doc.body().text());
        String code = o.getJSONObject("status").getString("code");
        if (StringUtils.equals(code, "1")) {
            return o.getJSONObject("domain").getString("id");
        }
        throw new DnspodException(o.getJSONObject("status").getString("message"));
    } catch (IOException e) {
        throw new DnspodException(e.getMessage());
    }
}

From source file:net.noday.core.dnspod.Dnspod.java

public static String domainInfo(String dnspodDomainId) {
    Document doc;
    try {/* w ww.ja  va 2s .  c  om*/
        doc = Jsoup.connect(url_domainInfo).data(data).data("domain_id", dnspodDomainId).userAgent(user_agent)
                .post();
        JSONObject o = JSON.parseObject(doc.body().text());
        String code = o.getJSONObject("status").getString("code");
        if (StringUtils.equals(code, "1")) {
            return o.getJSONObject("domain").getString("ext_status");
        }
        throw new DnspodException(o.getJSONObject("status").getString("message"));
    } catch (IOException e) {
        throw new DnspodException(e.getMessage());
    }
}

From source file:net.noday.core.dnspod.Dnspod.java

public static void domainRemove(String dnspodDomainId) {
    Document doc;
    try {//from w w  w  .j  a  v a2  s.  c om
        doc = Jsoup.connect(url_domainRemove).data(data).data("domain_id", dnspodDomainId).userAgent(user_agent)
                .post();
        JSONObject o = JSON.parseObject(doc.body().text());
        String code = o.getJSONObject("status").getString("code");
        if (!StringUtils.equals(code, "1")) {
            throw new DnspodException(o.getJSONObject("status").getString("message"));
        }
    } catch (IOException e) {
        throw new DnspodException(e.getMessage());
    }
}