Discord音乐机器人无法完整播放歌曲，请求排查代码问题

已确认的问题点 #

• FFmpeg重连配置未生效：代码中已经定义了带断流重连的ffmpeg_options参数，但初始化FFmpegPCMAudio实例时没有传入该参数，YouTube音频流存在访问时效，中途断流后没有重连逻辑就会直接停止播放
• 代码缩进错误：if not voice_client.is_playing()所在的代码块缩进错误，不属于play函数的作用域，实际运行时逻辑执行异常
• 异步方法调用错误：ctx.channel.send("I'm here!")是异步方法，没有添加await关键字会抛出未捕获的异步异常，可能导致进程中断、播放停止
• 音频流地址获取方式错误：直接取info["formats"][0]["url"]无法保证拿到稳定可用的最高音质音频流，容易拿到无效或者延迟很高的流地址
• 依赖库过时：youtube_dl已经停止维护多年，对新版YouTube的流解析适配很差，是流不稳定的核心诱因之一

修复后的完整代码 #

import discord
from discord.ext import commands
from discord.utils import get
# 建议将youtube_dl替换为yt-dlp，使用前执行pip install yt-dlp
from yt_dlp import YoutubeDL


bot = commands.Bot(command_prefix = ">", intents=discord.Intents.all())

@bot.event
async def on_ready():
    print("I'm ready for command")

@bot.command()
async def Help(ctx):
    emBed = discord.Embed(title = "How to use XXX Bot.", description = "This is every command you can use.", color = 0x03d7fc)
    emBed.add_field(name = ">Help" , value ="To see every command", inline = False)
    emBed.add_field(name = ">play" , value ="Play music", inline = False)
    emBed.add_field(name = ">pause" , value ="Pause music ", inline = False)
    emBed.add_field(name = ">skip" , value ="Stop music", inline = False)
    await ctx.channel.send(embed = emBed)

# 播放错误回调
async def play_error(error, ctx):
    if error:
        await ctx.send(f"播放出错：{str(error)}")

@bot.command()
async def play(ctx, url):
    channel = ctx.author.voice.channel
    voice_client = get(bot.voice_clients, guild = ctx.guild)
    
    if voice_client == None:
        await ctx.channel.send("I'm here!")
        await channel.connect()
        voice_client = get(bot.voice_clients, guild = ctx.guild)

    YDL_OPTIONS = {"format" : "bestaudio" , "noplaylist" : True, "quiet": True}
    ffmpeg_options = {"options": "-vn", "before_options": "-reconnect 1 -reconnect_streamed 1 -reconnect_delay_max 5" }

    if not voice_client.is_playing():
        with YoutubeDL(YDL_OPTIONS) as ydl:
            info = ydl.extract_info(url, download = False)
        # 直接取解析后的最优流地址
        stream_url = info["url"]
        # 传入ffmpeg配置，添加错误回调
        voice_client.play(discord.FFmpegPCMAudio(stream_url, **ffmpeg_options), after=lambda e: bot.loop.create_task(play_error(e, ctx)))
    else:
        await ctx.channel.send("I'm playing right now")
        return

bot.run("你的机器人TOKEN")

注意事项 #

• 运行前需确认已经安装了FFmpeg并配置到系统环境变量中
• 旧版discord.py需要开启意图权限，在Discord开发者后台的机器人设置页打开所有Privileged Gateway Intents开关
• 如果仍然使用原youtube_dl，只需将导入语句改回即可，但还是建议尽快替换为yt-dlp保证解析稳定性

内容的提问来源于stack exchange，提问作者Yuuuuuuu