python爬取QQ音乐评论信息

1.随便选个音乐

https://y.qq.com/n/yqq/song/0039MnYb0qxYhV.html

找到评论的数据接口：

https://c.y.qq.com/base/fcgi-bin/fcg_global_comment_h5.fcg?g_tk_new_20200303=1728726093&g_tk=1728726093&loginUin=0&hostUin=0&format=json&inCharset=utf8&outCharset=GB2312&notice=0&platform=yqq.json&needNewCode=0&cid=205360772&reqtype=2&biztype=1&topid=97773&cmd=8&needmusiccrit=0&pagenum=0&pagesize=25&lasthotcommentid=&domain=qq.com&ct=24&cv=10101010

参数列表：

参数名	参数值
g_tk_new_20200303	1728726093
g_tk	1728726093
loginUin	0
hostUin	0
format	json
inCharset	utf8
outCharset	GB2312
notice	0
platform	yqq.json
needNewCode	0
cid	205360772
reqtype	2
biztype	1
topid	97773
cmd	8
needmusiccrit	0
pagenum	0
pagesize	25
lasthotcommentid
domain	qq.com
ct	24
cv	10101010

可以发现两者值相等

参数名	参数值
g_tk_new_20200303	1728726093
g_tk	1728726093

2. g_tk函数介绍

在百度后才知道 g_tk 是个加密函数 ，而且还是javascript写的

然后再所有 js 中搜索 g_tk 按 Ctrl +F

终于找到了


d 引用的是 cookie.js

https://y.gtimg.cn/music/portal/js/common/music/cookie.js

代码:

define("js/common/music/cookie.js", function(e, t, n) {
	var o = e("js/common/config.js"),
		i = {
			set: function(e, t, n, i, c) {
				if (c) {
					var r = new Date;
					r.setTime(r.getTime() + 36e5 * c)
				}
				return document.cookie = e + "=" + escape(t) + "; " + (c ? "expires=" + r.toGMTString() + "; " : "") + (i ? "path=" + i + "; " : "path=/; ") + (n ? "domain=" + n + ";" : "domain=" + o.DCCookieDomain + ";"), !0
			},
			get: function(e) {
				var t;
				return function(e) {
					if (!e) return e;
					for (; e != unescape(e);) e = unescape(e);
					for (var t = ["<", ">", "'", '"', "%3c", "%3e", "%27", "%22", "%253c", "%253e", "%2527", "%2522"], n = ["&#x3c;", "&#x3e;", "&#x27;", "&#x22;", "%26%23x3c%3B", "%26%23x3e%3B", "%26%23x27%3B", "%26%23x22%3B", "%2526%2523x3c%253B", "%2526%2523x3e%253B", "%2526%2523x27%253B", "%2526%2523x22%253B"], o = 0; o < t.length; o++) e = e.replace(new RegExp(t[o], "gi"), n[o]);
					return e
				}((t = document.cookie.match(RegExp("(^|;\\s*)" + e + "=([^;]*)(;|$)"))) ? unescape(t[2]) : "")
			},
			del: function(e, t, n) {
				document.cookie = e + "=; expires=Mon, 26 Jul 1997 05:00:00 GMT; " + (n ? "path=" + n + "; " : "path=/; ") + (t ? "domain=" + t + ";" : "domain=" + o.DCCookieDomain + ";")
			},
			getACSRFToken: function(e, t) {
				e = e || "skey";
				var n = "",
					o = 5381;
				if (n = t ? i.get("qqmusic_key") || i.get("p_skey") || i.get("skey") || i.get("p_lskey") || i.get("lskey") : i.get(e) || i.get("skey") || i.get("p_skey") || i.get("p_lskey") || i.get("lskey") || i.get("qqmusic_key")) for (var c = 0, r = n.length; c < r; ++c) o += (o << 5) + n.charCodeAt(c);
				return 2147483647 & o
			}
		};
	return i
});

所以生成 g_tk 函数如下：

getACSRFToken: function(e, t) {
	e = e || "skey";
	var n = "",
		o = 5381;
	if (n = t ? i.get("qqmusic_key") || i.get("p_skey") || i.get("skey") || i.get("p_lskey") || i.get("lskey") : i.get(e) || i.get("skey") || i.get("p_skey") || i.get("p_lskey") || i.get("lskey") || i.get("qqmusic_key")) for (var c = 0, r = n.length; c < r; ++c) o += (o << 5) + n.charCodeAt(c);
	return 2147483647 & o
}

三元表达式：
表达式1？表达式2：表达式3
表达式1是一个条件，值为Boolean类型
若表达式1的值为true，则执行表达式2的操作，并且以表达式2的结果作为整个表达式的结果；
若表达式1的值为false，则执行表达式3的操作，并且以表达式3的结果作为整个表达式的结果；

|| 任意 一个条件成立 就为成立

用python替换

def g_tk(skey):
    hash = 5381
    length = len(skey)
    for i in range(length):
    	if i < length:
            hash += (hash<<5) + ord(skey[i])
    return 2147483647&hash

但是需要传入一个参数

假设 n = t

i.get("qqmusic_key")

从 cookie 中取值

我们 先找找 cookie

def g_tk(skey):
    hash = 5381
    length = len(skey)
    for i in range(length):
    	if i < length:
            hash += (hash<<5) + ord(skey[i])
    return 2147483647&hash

print(g_tk("Q_H_L_2OSlz160ei1C3usBln7CHpyKTimzESqdlHwVunbIFXiBSZjlxL5LKbmz25RIkA3"))

打印

这不刚好是 g_tk 参数的值吗

这个值登录后获取，会改变的

Q_H_L_2OSlz160ei1C3usBln7CHpyKTimzESqdlHwVunbIFXiBSZjlxL5LKbmz25RIkA3

3. topid

4. python 代码

import requests
from urllib import parse
import json
import pandas as pd
import os
import time

def g_tk(skey):
    hash = 5381
    length = len(skey)
    for i in range(length):
    	if i < length:
            hash += (hash<<5) + ord(skey[i])
    return 2147483647&hash

page = 0

df_all = pd.DataFrame()

while True:
	comment_url = "https://c.y.qq.com/base/fcgi-bin/fcg_global_comment_h5.fcg?"
	headers = {'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/76.0.3809.132 Safari/537.36'}
	params = {
	'g_tk_new_20200303':g_tk("Q_H_L_2Ou6-160eWT0kjGauzHMtXeJETzw9dz7sEuZ-3rx4M2Mt3qowToV8PY0bK_XLLE"),
	'g_tk':g_tk("Q_H_L_2Ou6-160eWT0kjGauzHMtXeJETzw9dz7sEuZ-3rx4M2Mt3qowToV8PY0bK_XLLE"),
	'loginUin':'0',
	'hostUin':'0',
	'format':'json',
	'inCharset':'utf8',
	'outCharset':'GB2312',
	'notice':'0',
	'platform':'yqq.json',
	'needNewCode':'0',
	'cid':'205360772',
	'reqtype':'2',
	'biztype':'1',
	'topid':'97773',
	'cmd':'8',
	'needmusiccrit':'0',
	'pagenum':'0',
	'pagesize':'25',
	'lasthotcommentid':'',	
	'domain':'qq.com',
	'ct':'24',
	'cv':'10101010'
	}

	params= parse.urlencode(params)
	url = comment_url + params
	response = requests.get(url,headers=headers)
	result = response.text
	comment_info = json.loads(result)

	topid = comment_info['topid']
	topic_name=comment_info['topic_name']

	comment = comment_info['comment']
	comment_total = comment['commenttotal']
	comment_list = comment['commentlist'] 
	page_total =int((comment_total-1)/25) + 1
	
	if page >=100 :
	    break
	    

	for i in comment_list:
	    
	    comment_id = i['commentid']
	    avatar_url =i['avatarurl']
	    nick=i['nick']
	    try:
	        content=i['rootcommentcontent']
	    except Exception as e:
	        content = ''
	        
	    comment_time=i['time']
	    timeArray  = time.localtime(comment_time)
	    comment_time = time.strftime("%Y年%m月%d日 %H:%M:%S", timeArray)
	    
	    praise_num=i['praisenum']
	    vip_icon=i['vipicon']
	    if vip_icon == '' :
	    	vip_icon = '未开通会员'
	    else:
	        vip_icon = vip_icon[-9:-4]    
	        
	    df = pd.DataFrame({
	        '评论ID':comment_id,
	        '头像链接':avatar_url,
	        '昵称':nick,
	        '评论内容':content,
	        '评论时间':comment_time,
	        '点赞数量':praise_num,
	        '等级图标':vip_icon
	    },index=[0])
	    df_all = df_all.append(df, ignore_index=True)
	    
	page = page + 1
	time.sleep(1)
	print("第"+str(page)+"页内容获取完毕")
	
df_all.to_excel(os.getcwd()+"\\"+topic_name+'_'+str(comment_total)+'_最新评论.xlsx',index = False)

5. 结果

总共4000 多页，延迟1秒

需要一个多小时才能跑完

暂时只演示 100条数据

if page >=100 :
    break

太长了 上传不了随意截图了一段