Migrate pulseaudio to pipewire (New)

[hanhsuan]

Description

1. using check_audio_deamon.sh to check what environment is under testing to use right function for testing
2. Under pipewire, using pw-dump and wpctl to replace pactl command to collect information and settings.
3. the output of pw-dump is json format, therefore pactl.py isn't needed a alternative for pipewire.
4. audio_settings.py is buggy for selecting right device, change to use show_default_audio_device.sh to show the default device and let user to check.
5. using wpctl status to replace audio_settings.py store command

Resolved issues

#327
After Ubuntu 22.04, the audio will change from pulseaudio to pipewire. Therefore, the old test case should be migrated from pulseaudio to pipewire.

JIRA

Documentation

Tests

Jira Card CQT-2396

[Hook25]

/canonical/self-hosted-runners/run-workflows 94ea095

[codecov]

Codecov Report

Attention: 47 lines in your changes are missing coverage. Please review.

Comparison is base (120bb96) 34.83% compared to head (10ceb84) 37.23%.
Report is 114 commits behind head on main.

Files	Patch %	Lines
providers/base/bin/pipewire_test.py	91.16%	14 Missing and 11 partials ⚠️
providers/base/bin/pipewire_utils.py	90.98%	9 Missing and 13 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #826      +/-   ##
==========================================
+ Coverage   34.83%   37.23%   +2.40%     
==========================================
  Files         302      304       +2     
  Lines       34165    34774     +609     
  Branches     5909     6000      +91     
==========================================
+ Hits        11901    12949    +1048     
+ Misses      21698    21226     -472     
- Partials      566      599      +33     

Flag	Coverage Δ
provider-base	10.19% <91.08%> (+7.06%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[diohe0311]

/canonical/self-hosted-runners/run-workflows a48b4b2

[diohe0311]

audio_settings.py is buggy for selecting right device

Can you tell more about this issue? in which env?

[hanhsuan]

For DP, HDMI and Thunderbolt monitor, the name shows in OS are the same with only index different. Therefore, audio_settings.py will never know which one is correct then force to select one that might not the right one. If the validator isn't noticed this issue, the validator might make the wrong judgement.

[diohe0311]

For DP, HDMI and Thunderbolt monitor, the name shows in OS are the same with only index different. Therefore, audio_settings.py will never know which one is correct then force to select one that might not the right one. If the validator isn't noticed this issue, the validator might make the wrong judgement.

Is this the only issue you have with audio_settings.py?
If audio_settings.py can provide more information, would you still need show_default_audio_device.sh?
It seems that the scripts have overlapping functionalities, perhaps instead of introducing a new script, we could work on improving the existing one?
@Hook25 @kissiel Do you have any suggestions?

[hanhsuan]

Other functions depend on pulseaudio that couldn't work on pipewire.

[diohe0311]

Other functions depend on pulseaudio that couldn't work on pipewire.

I'm not familiar with PulseAudio and Pipewire, just start to learning it because of this PR :)
I found Pipewire has a pipewire-pulse.service, it supports PulseAudio applications.
https://wiki.archlinux.org/title/PipeWire

We can discuss more once I figure out how to use it.

[hanhsuan]

1. pactl is not installed by default after lunar
2. The output format of pactl info is different between pulseaudio and pipewire with pipewire-pulse, it makes the pactl.py parsing failed.

[Hook25]

Allow me to endorse this comment: #826 (comment)

Looking at the changes from audio_test.py -> pipewire_test.py, it looks very clear to me that the two should be merged and use inheritance to obtain the result that we need.

Let me be more explicit, we have a super class here that is VolumeController, this class should contain all common functionality and the interface.

def run_retry(command):
    # Since the audio system may return a failure if the audio layer is not 
    # yet initialized, we will try running a few times in case of failure. 
    # Do not use this if your command is not "idempotent", 
    # repeating them should not have any bad effects.
    for attempt in range(0, 3):
        try:
            return subprocess.check_output(
                command,
                universal_newlines=True
            )
        except (subprocess.CalledProcessError):
            time.sleep(5)

class VolumeController:
    def __init__(self, controller_type, logger=None):
        self.controller_type = controller_type
        self.logger = logger
    def set_volume(self, volume): pass
    def get_volume(self): pass
    def set_mute(self, mute): pass

Then we have two subclasses, one for pipewire and one for pulseaudio

class PulseVolumeController(VolumeController):
    ... # this is the one from audio_test.py with minor edits

class PipewireVolumeController(VolumeController):
    ... # this is from pipewire_test.py

The main function is basically the same for the two modules (but the warning message, that should be moved inside the PulseVolumeController class. To select the audio system we want to test we can add a new cli argument with two possible choiches (pipewire, pulseaudio), and we initialize the correct instance in main depending on this argument value.

This kind of design should also make it trivial to add a new audio system or remove an old one, for example, if we ever deprecate testing pulseaudio systems via checkbox, we can just remove the argument and the class, and everything else will remain the same.

I did read most of this PR but I feel like the comments by @diohe0311 are correct, once those are addressed I can give some additional comments (I don't want them to overlap, lets do one review at the time!)

[diohe0311]

@hanhsuan Hi! we have updated the code review criteria

I changed my formatting related comments to suggestions, @Hook25 Hanhsuan mentioned you guys will have a review meeting for this, just want to let you know that the mock input in test_pipewire_utils.py is highly identical(some of them is the same), but I'm not sure if it worth to do something to simplify it or just leave it be.

[Hook25]

+1, this is probably going to need a followup but its a good base to work on. Thank you

Fixed by the latest commits