public static void main(String args[]) throws Exception {
HtmlObject htmlObject = new HtmlObject();
StringBuffer sb = new StringBuffer();
BufferedReader reader = new BufferedReader(new FileReader("/home/chen/example/cifNews.html"));
String temp = null;
while ((temp = reader.readLine()) != null) {
sb.append(temp);
}
htmlObject.setHtml(sb.toString());
htmlObject.setUrl("http://www.cifnews.com/Article/14175");
CifNewsParser parser = new CifNewsParser();
Article article = parser.run(htmlObject);
System.out.println("title:" + article.getTitle());
System.out.println("content:" + article.getContent());
System.out.println("date:" + article.getPublishDate());
}
@Override
public Article run(HtmlObject htmlObject) {
String html = htmlObject.getHtml();
Document doc = Jsoup.parse(html);
String title = doc.select(".article h1").text();
Elements contentElement = doc.select(".article_con");
String content = "";
String contentHtml = "";
if (contentElement != null) {
// contentElement.select(".author").remove();
content = contentElement.text();
contentHtml = contentElement.html();
}
String Ele_data = doc.select(".article h2").text();
Matcher m1 = datePattern.matcher(Ele_data);
String date = "";
if (m1.find()) {
date = m1.group(1);
} else {
Date today = new Date();
SimpleDateFormat formatter = new SimpleDateFormat("yyyy-MM-dd");
date = formatter.format(today);
}
Article model1 = new Article();
model1.setUrl(htmlObject.getUrl());
model1.setTitle(title);
model1.setContent(content);
model1.setPublishDate(date);
model1.setArticleType(ArticleType.News);
model1.setProvider("雨果网");
return model1;
}