python笔记(爬虫 微爬取微信信息)

views.py

import time
import json
import re
import requests
from bs4 import BeautifulSoup
from flask import Blueprint, render_template, session, jsonify, request


wx = Blueprint("wx", __name__)


def xml_parse(text):
    result = {}
    soup = BeautifulSoup(text, "html.parser")
    tag_list = soup.find(name="error").find_all()
    for tag in tag_list:
        result[tag.name] = tag.text
    return result


@wx.route("/login")
def login():
    ctime = int(time.time() * 1000)
    wcode_url = "https://login.wx.qq.com/jslogin?appid=wx782c26e4c19acffb&redirect_uri=https%3A%2F%2Fwx.qq.com%2Fcgi-bin%2Fmmwebwx-bin%2Fwebwxnewloginpage&fun=new&lang=zh_CN&_={}".format(ctime)
    res = requests.get(
        url=wcode_url
    )
    data = res.text
    uuid = re.findall('uuid = "(.*)";', data)[0]
    session["uuid"] = uuid
    return render_template("wechat_page.html", uuid=uuid)


@wx.route("/check_login")
def check_login():
    uuid = session.get("uuid")
    ctime = int(time.time() * 1000)
    check_login_url = "https://login.wx.qq.com/cgi-bin/mmwebwx-bin/login?loginicon=true&uuid={}&tip=0&r=-1916563215&_={}".format(uuid, ctime)
    res = requests.get(
        url=check_login_url
    )
    result = {"code": None}
    if "window.code=408" in res.text:
        result["code"] = 408
    elif "window.code=201" in res.text:
        result["code"] = 201
        result["userAvatar"] = re.findall("window.userAvatar = '(.*)';", res.text)[0]
    elif "window.code=200" in res.text:
        result["code"] = 200
        redirect_uri = re.findall('redirect_uri="(.*)";', res.text)[0]+"&fun=new&version=v2"
        ret = requests.get(
            url=redirect_uri
        )
        ticket_dict = xml_parse(ret.text)
        session["ticket_dict"] = ticket_dict
        session["ticket_cookie"] = ret.cookies.get_dict()
    return jsonify(result)


@wx.route("/mainpage")
def mainpage():
    pass_ticket = session["ticket_dict"].get("pass_ticket")
    init_url = " https://wx2.qq.com/cgi-bin/mmwebwx-bin/webwxinit?r=-1932693400&pass_ticket={}".format(pass_ticket)
    res = requests.post(
        url=init_url,
        json={
            "BaseRequest": {
                "DeviceID": "e145195217922834",
                "Sid": session["ticket_dict"].get("wxsid"),
                "Skey": session["ticket_dict"].get("skey"),
                "Uin": session["ticket_dict"].get("wxuin")
            }
        }
    )
    res.encoding = "utf-8"

    init_user_dict = res.json()
    return render_template("wechat_index.html", init_user_dict=init_user_dict)


@wx.route("/contractlist")
def contractlist():
    ctime = int(time.time() * 1000)
    # pass_ticket = session["ticket_dict"].get("pass_ticket")
    skey = session["ticket_dict"].get("skey")
    print(session.get("ticket_cookie"))
    contract_url = "https://wx2.qq.com/cgi-bin/mmwebwx-bin/webwxgetcontact?r={0}&seq=0&skey={1}".format(ctime, skey)
    res = requests.get(
        url=contract_url,
        cookies=session.get("ticket_cookie")
    )
    res.encoding = "utf-8"
    user_list = res.json()
    return render_template("contact_list.html", userlist=user_list.get("MemberList"))


@wx.route("/get_img")
def get_img():
    prev = request.args.get("prev")
    username = request.args.get("username")
    skey = request.args.get("skey")
    head_img_url = "https://wx2.qq.com{0}&username={1}&skey={2}".format(prev, username, skey)
    res = requests.get(
        url=head_img_url,
        cookies=session.get("ticket_cookie")
    )
    return res.content


@wx.route("/send_msg", methods=["POST"])
def send_msg():
    ctime = int(time.time() * 1000)
    from_user = request.form.get("fromuser")
    to_user = request.form.get("touser")
    content = request.form.get("content")

    data_dict = {
        "BaseRequest": {
            "DeviceID": "e823270257010524",
            "Sid": session["ticket_dict"].get("wxsid"),
            "Skey": session["ticket_dict"].get("skey"),
            "Uin": session["ticket_dict"].get("wxuin")
        },
        "Msg": {
            "ClientMsgId": ctime,
            "Content": content,
            "FromUserName": from_user,
            "LocalID": ctime,
            "ToUserName": to_user,
            'Type': 1
        },
        'Scene': 0
    }
    send_url = "https://wx2.qq.com/cgi-bin/mmwebwx-bin/webwxsendmsg?lang=zh_CN&pass_ticket={0}".format(session["ticket_dict"].get("pass_ticket"))
    ret = requests.post(
        url=send_url,
        data=bytes(json.dumps(data_dict, ensure_ascii=False), encoding="utf-8")
    )
    print(ret)
    return "发送成功"

wechat_page.html

<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <title>微信</title>
</head>
<body>
<div>
    <h1>扫码登录</h1>
    <img id="userAvatar" src="https://login.weixin.qq.com/qrcode/{{uuid}}" alt="">
</div>

<script src="https://cdn.bootcss.com/jquery/3.4.0/jquery.min.js"></script>
<script>
$(function () {
    check_login();
});
function check_login() {
    $.ajax({
        url:"/check_login",
        method:"GET",
        dataType:"json",
        success:function (arg) {
            if (arg.code === 408){
               check_login();
            }
            else if (arg.code === 201){
                $("#userAvatar").attr("src", arg.userAvatar);
                check_login();
            }
            else if (arg.code === 200){
                location.href = "/mainpage"
            }
        }
    })
}
</script>
</body>
</html>

wechat_index.html

<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <title>微信首页</title>
</head>
<body>
<h1>欢迎来到微信{{init_user_dict.User.UserName}}</h1>

<h3>最近联系人</h3>
<ul>
    {% for row in init_user_dict.ContactList %}
    <li>{{row.NickName}}</li>
    {% endfor %}
    <a href="/contractlist">联系人列表</a>
</ul>
<h3>最近公众号</h3>
{% for item in init_user_dict.MPSubscribeMsgList %}
<div>
    <h3>{{item.NickName}}</h3>
    <ul>
        {% for msg in item.MPArticleList %}
        <li><a href="{{msg.Url}}">{{msg.Title}}</a></li>
        {% endfor %}
    </ul>
</div>
{% endfor %}
</body>
</html>

contact_list.html

<!DOCTYPE html>
<html lang="en">
<head>
    <meta charset="UTF-8">
    <title>联系人列表</title>
</head>
<body>
<div>
    <form action="/send_msg" method="post">
        <p>
            自己的唯一标识：<input type="text" name="fromuser">
            别人的唯一标识：<input type="text" name="touser">
            发送内容：<input type="text" name="content">
            <button type="submit">发送</button>
        </p>
    </form>
</div>

<h1>联系人列表</h1>
{% for user in userlist %}
<li>
     <img style="height: 50px;width: 50px;" src="/get_img?prev={{user.HeadImgUrl}}">
    <span>用户名：{{user.NickName}} 唯一标识：{{user.UserName}}</span>
</li>
{% endfor %}
</body>
</html>