Scrobbling to youtube music not reflected in Recent History

[kmac]

The scrobbling code seems to be working as implemented, however nothing shows up in Recent History.

We get a 204 from /api/stats/playback:

DEBUG    2022-05-16 14:06:56,735 [1822438:YTMusicBackend-7 (_actor_loop)] mopidy_ytmusic 204 code from
'https://s.youtube.com/api/stats/playback?cl=...&docid=...&ei=...&fexp=...&ns=y│t&list=...&el=detailpage&len=643&of=-...&cpn=.&
referrer=https%3A%2F%2Fmusic.youtube.com&ver=2'

I suppose it's possible the API/mechanism changed for updating? Are others seeing the same thing?

[dionorgua]

As far as I see it's just not implemented:
https://github.com/OzymandiasTheGreat/mopidy-ytmusic/blob/master/mopidy_ytmusic/scrobble_fe.py

[impliedchaos]

@kmac - A 204 is probably what we should get back. I've been swamped with work and haven't had a chance to look at this. I'll see if I can soon.

@dionorgua - the pass in the function lets the function in backend.py handle it. It's a weirdly confusing way to do it, but it's how it was done in the Google Play Music backend that I was stealing logic from, so I left it. 😅
https://github.com/OzymandiasTheGreat/mopidy-ytmusic/blob/master/mopidy_ytmusic/backend.py#L161

[impliedchaos]

Also, if I do manage to get it working again (or someone beats me to it 🙏), we should probably discuss with @sigma67 to put it in ytmusicapi instead. There're other history manipulation functions in there already.

[sigma67]

Very much open to a PR, this one's just flown under the radar a bit.

[impliedchaos]

This should finally be fixed in 0.3.8

[sigma67]

So I've looked at the code and thought about how to implement this in ytmusicapi, but it seems it's pretty much identical to get_song. The only thing that's missing is the CPN in the request?

[impliedchaos]

The CPN doesn't even matter for the player API call (just the tracking URL call). I only wrote it like this so I could get the playbackTracking object that get_song skips. If it was added to the list of keys in get_song, then this could be much simpler like:

import random
import requests
from ytmusicapi import YTMusic

YTMAPI = YTMusic("auth.json")

videoId = "xZdj3LM5X1M"  # Lizzo - Good as Hell
song = YTMAPI.get_song(videoId)

CPNA = ("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789-_")
cpn = "".join( ( CPNA[random.randint(0, 256) & 63] for _ in range(0, 16) ) )
response = requests.get(
    song["playbackTracking"]["videostatsPlaybackUrl"]["baseUrl"],
    params={
        "ver": 2,
        "c": "WEB_REMIX",
        "cpn": cpn,
    },
    headers=YTMAPI.headers,
    proxies=YTMAPI.proxies,
)
if (response.status_code == 204):
    print("\"%s\" added to history." % song["videoDetails"]["title"])

And the correct signatureTimestamp doesn't even need to be used to get back the correct urls in playbackTracking.

[sigma67]

Ah alright, that's a very simple change. There's no reason to exclude that key apart from excluding non-useful data. Which playbackTracking obviously isn't as you just provided a use-case. Will implement shortly.