@Override
public void process(ResultItems resultItems, Task task) {
System.out.println("get page: " + resultItems.getRequest().getUrl());
for (Map.Entry<String, Object> entry : resultItems.getAll().entrySet()) {
System.out.println(entry.getKey() + ":\t" + entry.getValue());
}
}
@Override
public void process(ResultItems items, Task task) {
// TODO Auto-generated method stub
System.out.println("对应的页面: " + items.getRequest().getUrl());
/*
* 抓取到的name和id
* */
List<String> name = new ArrayList<String>();
List<String> aid = new ArrayList<String>();
List<String> url = new ArrayList<String>();
FileWriter fw = null;
FileWriter fwPlayUrl = null;
for (Map.Entry<String, Object> entry : items.getAll().entrySet()) {
String key = (String) entry.getKey();
if (key.equals(CrawlerIqiyi.VIDEO_NAME)) {
name = (List<String>) entry.getValue();
} else if (key.equals(CrawlerIqiyi.VIDEO_ID)) {
aid = (List<String>) entry.getValue();
} else if (key.equals(CrawlerIqiyi.VIDEO_URL)) {
url = (List<String>) entry.getValue();
}
}
try {
fw = new FileWriter("./urls.txt", true);
fwPlayUrl = new FileWriter("./playUrls.txt", true);
for (int i = 0; i < name.size(); i++) {
/*
System.out.println(name.get(i)+" 指数页面为: "+
"http://index.iqiyi.com/q/?aid="+aid.get(i)+"&name="
+ name.get(i));*/
fw.write(
"http://uaa.iqiyi.com/video_index/v1/get_user_profile?album_id="
+ aid.get(i)
+ "&album_name="
+ name.get(i)
+ "&callback=window.Q.__callbacks__.cbgg1cdr");
fw.write("\n");
fwPlayUrl.write(url.get(i));
fwPlayUrl.write("\n");
}
fw.close();
fwPlayUrl.close();
} catch (Exception e) {
e.printStackTrace();
}
}
public static void main(String[] args) {
// single download
Spider spider = Spider.create(new BaiduBaikePageProcessor()).thread(2);
String urlTemplate = "http://baike.baidu.com/search/word?word=%s&pic=1&sug=1&enc=utf8";
ResultItems resultItems = spider.<ResultItems>get(String.format(urlTemplate, "水力发电"));
System.out.println(resultItems);
// multidownload
List<String> list = new ArrayList<String>();
list.add(String.format(urlTemplate, "风力发电"));
list.add(String.format(urlTemplate, "太阳能"));
list.add(String.format(urlTemplate, "地热发电"));
list.add(String.format(urlTemplate, "地热发电"));
List<ResultItems> resultItemses = spider.<ResultItems>getAll(list);
for (ResultItems resultItemse : resultItemses) {
System.out.println(resultItemse.getAll());
}
spider.close();
}
@Override
public void process(ResultItems resultItems, Task task) {
// TODO Auto-generated method stub
List<Good> goodList = resultItems.get("goodList");
dao.addGoods(goodList);
}
