cosven / feeluown-core

提供音乐播放器的一些常见组成模块（已经合并到 feeluown 项目中）

feeluown-core

fuocore 提供了音乐播放器依赖的一些常见模块， 它主要是为 feeluown 播放器而设计的， 所以名字为 fuocore，意为 feeluown core。理论上其它音乐播放器也可以使用 fuocore 作为其基础模块。

👉 详细文档

feeluown-core's Issues

不能自动播放下一曲

问题描述：

一首歌播放结束之后，不能自动播放下一首。

Debug

多线程编程经验不足
https://github.com/cosven/feeluown-core/blob/master/fuocore/player.py#L124 这句会报：can't join current thread.

解决思路

1. Signal 跨线程，像 Qt 那样

AttributeError: module 'logging' has no attribute 'config'

运行feel-core-test，报错"AttributeError: module 'logging' has no
attribute 'config'"。在fuocore/main.py中添加
import logging.config
后，错误消失。

将一首歌从播放列表移除后，播放下一首歌就会报错

[ERROR] [__init__ exec_cmd 75] : handle cmd(action:status args:()) error
Traceback (most recent call last):
  File "/home/cosven/.local/lib/python3.5/site-packages/fuocore/protocol/handlers/__init__.py", line 71, in exec_cmd
    cmd_rv = handler.handle(cmd)
  File "/home/cosven/.local/lib/python3.5/site-packages/fuocore/protocol/handlers/__init__.py", line 119, in handle
    'song:      {}'.format(show_song(player.current_song, brief=True)),  # noqa
  File "/home/cosven/.local/lib/python3.5/site-packages/fuocore/protocol/handlers/helpers.py", line 15, in show_song
    artists = song.artists or []
AttributeError: 'NoneType' object has no attribute 'artists'

当扫描到一些不是视频文件的时候出错

我有一些MV视频在歌曲文件夹里面。当运行 fuo --debug的时候报错：

[INFO] [__main__ main 17] : Release mode.
Traceback (most recent call last):
  File "/home/stardiviner/.virtualenvs/python3.6/bin/fuo", line 11, in <module>
    sys.exit(main())
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/fuocore/__main__.py", line 20, in main
    local_provider = LocalProvider()
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/fuocore/local/provider.py", line 48, in __init__
    self.library_paths = library_paths
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/fuocore/local/provider.py", line 57, in library_paths
    self._songs = self.scan()
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/fuocore/decorators.py", line 13, in wrapper
    result = func(*args, **kwargs)
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/fuocore/local/provider.py", line 84, in scan
    songs = [self.model_from_fpath(fpath) for fpath in media_files]
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/fuocore/local/provider.py", line 84, in <listcomp>
    songs = [self.model_from_fpath(fpath) for fpath in media_files]
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/fuocore/local/provider.py", line 93, in model_from_fpath
    metadata = EasyMP3(fpath)
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/mutagen/_file.py", line 49, in __init__
    self.load(*args, **kwargs)
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/mutagen/_util.py", line 140, in wrapper
    return func(self, h, *args, **kwargs)
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/mutagen/id3/_file.py", line 406, in load
    self.info = self._Info(fileobj, offset)
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/mutagen/_util.py", line 169, in wrapper
    return func(*args, **kwargs)
  File "/home/stardiviner/.virtualenvs/python3.6/lib/python3.6/site-packages/mutagen/mp3/__init__.py", line 403, in __init__
    raise HeaderNotFoundError("can't sync to MPEG frame")
mutagen.mp3.HeaderNotFoundError: can't sync to MPEG frame

music player DSL?

音乐播放有什么操作

以前的磁带、复读机：快退，暂停/播放，停止，快进
MP3：复读机的所有功能，曲库管理，选择某一首播放，上一首，下一首；

> resume 
> pause
> stop
> play 晴天
> seek 2:34

改进player.playlist的数据结构

player.playlist 里使用list来存放歌曲, 在查询/添加/删除歌曲的时候复杂度均为O(n). 如果想要支持衍生的一些想法中几万张专辑这样的大歌单, 应该会非常低效.

目前有两个想法:

1. linked list: 利用一个链表来存放歌曲, 然后用一个dict索引 song->linked list node, 从而维护歌单的有序结构.
2. list with empty slots: 同样利用list和dict存放歌曲, 删除歌曲的时候设为None保留空位, 然后当歌曲小于2/3 list长度的时候resize.

3种方案比较:

operation	raw list	linked list	list with empty slots
append	O(N)	O(1)	O(1)
search	O(N)	O(1)	O(1)
remove	O(N)	O(1)	Amortized O(1)
memory	song size	song size + dict size + references x N	song size + dict size
drawback	too slow	can't directly access by index	insertion after a song is O(N)

目前利用整数代替歌曲测试, 添加100000整数, 然后随机删除:

• raw list: append 95.6s, remove 70.2s ;
• linked list: append 0.35s, remove 0.26s;
• list with empty slots: append 0.05s, remove 0.12s.

可以看到第三种方案相当迅速(linked list需要在nodes 之间跳转), 但缺点在于插入歌曲时的效率依然较差.

另外, 我觉得在playlist里不必要维护 good songs和bad songs, 在 player播放有误的时候跳过就好了.

声卡选择

这个可以选择播放的声卡吗？怎么操作呢？

TODO list.

• bugfix Restart failed, "Address already in use"
• refactor code structure