最新版某易云音乐的爬取（音乐+全部评论）

好久没写文章了，今天有时间更新一下吧，稍微看了一下某易云最新的JS加密：

首先是歌曲：先过滤出

/player/url/v1

params和encSecKey是两个关键字段，那么我们先来打断点找到这个加密混淆起始的地方，

控制台看一下果然是是这里

分析了一下后脱机算法写出来：

const crypto = require("crypto");

const presetKey = "0CoJUm6Qyw8W8jud";
const iv = "0102030405060708";
const pubKey = "010001";
const modulus =
  "00e0b509f6259df8642dbc35662901477df22677ec152b5ff68ace615bb7b725" +
  "152b3ab17a876aea8a5aa76d2e417629ec4ee341f56135fccf695280104e0312" +
  "ecbda92557c93870114af6c9d05c4f7f0c3685b7a46bee255932575cce10b424" +
  "d813cfe4875d3e82047b97ddef52741d546b8e289dc6935b3ece0462db0a22b8" +
  "e7";

function aesEncrypt(text, key) {
  const cipher = crypto.createCipheriv("aes-128-cbc", key, iv);
  cipher.setAutoPadding(true);
  return cipher.update(text, "utf8", "base64") + cipher.final("base64");
}

function rsaEncrypt(text) {
  text = text.split("").reverse().join("");
  const hexText = Buffer.from(text).toString("hex");
  const bigIntText = BigInt("0x" + hexText);
  const bigIntPubKey = BigInt("0x" + pubKey);
  const bigIntModulus = BigInt("0x" + modulus);

  const encrypted = bigIntText ** bigIntPubKey % bigIntModulus;
  return encrypted.toString(16).padStart(256, "0");
}

function createSecretKey(size) {
  const chars = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789";
  let key = "";
  for (let i = 0; i < size; i++) {
    key += chars.charAt(Math.floor(Math.random() * chars.length));
  }
  return key;
}

function weapi(data) {
  const text = JSON.stringify(data);
  const secKey = createSecretKey(16);

  const firstPass = aesEncrypt(text, presetKey);
  const params = aesEncrypt(firstPass, secKey);
  const encSecKey = rsaEncrypt(secKey);

  return { params, encSecKey };
}

// 主入口：命令行调用时返回 JSON
if (require.main === module) {
  const input = process.argv[2]; // 接收 JSON 字符串
  const data = JSON.parse(input);
  console.log(JSON.stringify(weapi(data)));
}

module.exports = weapi;

再写一个程序调用这段js：

import subprocess
import json
import requests

# ====== 调用 Node.js weapi.js 脚本 ======
def get_weapi_params(data: dict):
    process = subprocess.run(
        ["node", "weapi.js", json.dumps(data, ensure_ascii=False)],
        capture_output=True,
        text=True,
        check=True
    )
    return json.loads(process.stdout)

# ====== 构造 payload（关键！） ======
def build_song_payload(song_id: int, csrf: str = "") -> dict:
    return {
        "ids": f"[{song_id}]",
        "level": "exhigh",      
        "encodeType": "aac",
        "csrf_token": csrf
    }

# ====== 发请求 ======
def get_song_url(song_id: int):
    csrf_token = "替换成你自己的"

    # 1. 构造 payload
    payload = build_song_payload(song_id, csrf=csrf_token)

    # 2. weapi 加密
    crypto_data = get_weapi_params(payload)

    cookies = {
        "__csrf": csrf_token,
        "MUSIC_U": "替换成你自己的"
    }

    headers = {
        "Content-Type": "application/x-www-form-urlencoded",
        "Referer": "https://music.163.com/",
        "Origin": "https://music.163.com",
        "User-Agent": "Mozilla/5.0"
    }

    # 3. 请求接口
    resp = requests.post(
        "https://music.163.com/weapi/song/enhance/player/url/v1",
        data=crypto_data,
        headers=headers,
        cookies=cookies,
        params={"csrf_token": csrf_token}
    )

    try:
        return resp.json()
    except Exception:
        print(resp.text)
        return None


# ====== 示例调用（数字是歌曲ID） ======
print(get_song_url(这里添加数字))

也是能跑的：

点击连接也是听到了《罗生门》，这里我只是演示就不批量了。

全部评论：

这里我要说下，歌曲的params是216位，评论的params是364位。这是因为传入的参数不同。

废话不多说直接上代码：

import subprocess
import json
import requests
import time


# 调用 js 脚本生成 params + encSecKey
def get_weapi_params(data: dict):
    process = subprocess.run(
        ["node", "weapi.js", json.dumps(data)],
        capture_output=True,
        text=True,
        check=True
    )
    return json.loads(process.stdout)

def get_nickname(comment):
    user = comment.get("user") or {}
    return user.get("nickname") or user.get("name")


def get_comments(song_id, cursor, page_size, page_no):
    url = "https://music.163.com/weapi/comment/resource/comments/get"

    data = {
        "rid": f"R_SO_4_{song_id}",
        "threadId": f"R_SO_4_{song_id}",
        "pageNo": page_no,  # 页码
        "pageSize": page_size,
        "cursor": cursor,  
        "offset": (page_no - 1) * page_size,  
        "orderType": 1,
        "csrf_token": "替换你自己的",
    }

    crypto_data = get_weapi_params(data)

    headers = {
        "Content-Type": "application/x-www-form-urlencoded",
        "Referer": "https://music.163.com/",
        "Origin": "https://music.163.com",
        "User-Agent": "Mozilla/5.0"
    }

    cookies = {
        'MUSIC_U': '替换你自己的',
        '__csrf': '替换你自己的',
   }
    resp = requests.post(url, data=crypto_data, headers=headers, cookies=cookies)
    return resp.json(), crypto_data

    resp = requests.post(url, data=crypto_data, headers=headers)
    return resp.json(), crypto_data


if __name__ == "__main__":
    song_id = 1456890009   #歌曲ID
    cursor = -1
    page_size = 20
    page_no = 1

    for _ in range(5):
        result, crypto_data = get_comments(song_id, cursor, page_size, page_no)

        print(f"\n=== 第 {page_no} 页 ===")
        print(f"cursor={cursor}, params length={len(crypto_data['params'])}")

        comments = result.get("data", {}).get("comments", [])
        if not comments:
            print("没有评论了。")
            break

        # 只打印想要的字段
        for c in comments:
            location = c.get("ipLocation", {}).get("location")
            name = get_nickname(c)
            print(f"{c.get('commentId')} | {name} | {c.get('content')} | {c.get('timeStr')} | {location}")

        # 更新 cursor
        cursor = result.get("data", {}).get("cursor")
        page_no += 1
        time.sleep(1)

 

这里我只跑5页：

这个cursor就很有意思第一次请求是-1而后面会根据第一次返回的时间戳来请求：

第一次：

第二次：