[转载]爬虫-微信公众平台消息获取 - 推酷

[转载]爬虫-微信公众平台消息获取 – 推酷.

帮朋友抓取微信公众平台的用户评论信息。

下面只说核心的部分，怎么获取评论信息。

查看HTML代码，没有发现关于评论部分的标签。看来是用JS动态生成的，但是查找ajax请求也没有找到哪里有返回数据。

最后搜索一下，原来是在这里，很直白的写在了JS里：

<script type="text/javascript">
    wx.cgiData = {
      total_count : 91,
      latest_msg_id : '200325222',
      count : "20"*1 || 20,
      day : "7",
      frommsgid : "",
      can_search_msg : "1",
      offset : "",
      action : "",
      keyword : "",
      list : ({"msg_item":[{"id":200322761,"type":1,"fakeid":"593656935","nick_name":"Suang 1","date_time":1398854675,"content":"记得帮我查一下是不是这个电话！","source":"","msg_status":4,"has_reply":0,"refuse_reason":"","multi_item":[],"to_uin":3071594631,"send_stat":{"total":0,"succ":0,"fail":0}},{"id":200322760,"type":2,"fakeid":"593656935","nick_name":"Suang 1","date_time":1398854664,"source":"","msg_status":4,"has_reply":0,"refuse_reason":"","multi_item":[],"to_uin":3071594631,"send_stat":{"total":0,"succ":0,"fail":0}},{"id":200322759,"type":1,"fakeid":"593656935","nick_name":"Suang 1","date_time":1398854659,"content":"勐璇，我看到那人了！","source":"","msg_status":4,"has_reply":0,"refuse_reason":"","multi_item":[],"to_uin":3071594631,"send_stat":{"total":0,"succ":0,"fail":0}},{"id":200322344,"type":2,"fakeid":"1994400010","nick_name":"ABC的CBA","date_time":1398839849,"source":"","msg_status":4,"has_reply":0,"refuse_reason":"","multi_item":[],"to_uin":3071594631,"send_stat":{"total":0,"succ":0,"fail":0}},{"id":200321209,"type":1,"fakeid":"1591078101","nick_name":"倚(纺织服装)","date_time":1398788906,"content":"\/::<","source":"","msg_status":4,"has_reply":0,"refuse_reason":"","multi_item":[],"to_uin":3071594631,"send_stat":{"total":0,"succ":0,"fail":0}},{"id":200321206,"type":2,"fakeid":"1591078101","nick_name":"倚(纺织服装)","date_time":1398788859,"source":"","msg_status":4,"has_reply":1,"refuse_reason":"","multi_item":[],"to_uin":3071594631,"send_stat":{"total":0,"succ":0,"fail":0}},

用的是JSON格式，代码太乱，放在Eclipse里格式化一下，消息列表大概就是这个样了：

{"msg_item" :[ {
  "id" : 200322761,
  "type" : 1,
  "fakeid" : "593656935",
  "nick_name" : "Suang 1",
  "date_time" : 1398854675,
  "content" : "记得帮我查一下是不是这个电话！",
  "source" : "",
  "msg_status" : 4,
  "has_reply" : 0,
  "refuse_reason" : "",
  "multi_item" : [],
  "to_uin" : 3071594631,
  "send_stat" : {
    "total" : 0,
    "succ" : 0,
    "fail" : 0
  }
}, {
  "id" : 200322760,
  "type" : 2,
  "fakeid" : "593656935",
  "nick_name" : "Suang 1",
  "date_time" : 1398854664,
  "source" : "",
  "msg_status" : 4,
  "has_reply" : 0,
  "refuse_reason" : "",
  "multi_item" : [],
  "to_uin" : 3071594631,
  "send_stat" : {
    "total" : 0,
    "succ" : 0,
    "fail" : 0
  }
}
]
}

上面就是 json字符串 中 msg_item 所对应的列表里的对象。

可以看出这个是一个数组，每个评论是里面的一个对象。怎么生成对于的Java类呢 ？

这里有一个在线的工具：http://jsongen.byingtondesign.com/

可以根据JSON 字符串，生成对应的java类：

类1

import java.util.List;

public class MessageList{
  private List<Message> msg_item;

  public List<Message> getMsg_item() {
    return msg_item;
  }

  public void setMsg_item(List<Message> msgItem) {
    msg_item = msgItem;
  }
  
}

类2。部分字段没有用，删掉了

public class Message {

  private String content;
  private long date_time;
  private String fakeid;
  private int has_reply;
  private long id;
  private int msg_status;
  private String nick_name;
  private String refuse_reason;
  private String source;
  private long to_uin;
  private int type;
// get set 略去
}

下面来做个测试。用google的 Gson 来进行处理，把json字符串解析为 java对象。

//jsonstr 为 msg_item 的json字符串
    MessageList msgList = new Gson().fromJson(jsonstr, MessageList.class);
    System.out.println(msgList.getMsg_item().size());

解析成功。所有的对象都在 msgList里了