Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: foo_softplaylists (Read 244036 times) previous topic - next topic
0 Members and 2 Guests are viewing this topic.


foo_softplaylists

Reply #327
<error code="6">Invalid limit. Please supply a limit between 1 and 200</error></lfm>

So 200 it will be

foo_softplaylists

Reply #328
i notice last.fm still let you go wild on the loved tracks feed. this lack of consistency is quite annoying really. i'd be happy with 200 on all of them but some are supposedly limited to 50 still.



foo_softplaylists

Reply #331
Hello JohanDeBock,
I love you plugin. Really! Thanks!

Especially in this household with a big NAS, where all the people put their music into the foobar shares, I can enjoy music i would have never listened to while searching the media library by hand.

Great! :-)

but anyway...

A request:

If i choose "Similar Tracks", your fine plugin has the option to "ABORT" the search for similar tracks.
But if i press "ABORT", your plugin opens a new playlist with no music in it.

I'd rather like to have all those music in the list which your plugin was able to find before i pressed "ABORT".

so, if the plugin searches for more than 5 minutes for similar tracks on last.fm, and i don't want to wait anymore, i'd like to press a button,  like i.e. "STOP THE CURRENT SEARCH AND START PLAYING THE UNCOMPLETED LIST IMEDIATELY"

is this possible?
would you mind to implement such feature, if possible?

cheers,



foo_softplaylists

Reply #332
Just found a bug with the latest version, 2011-02-05, using foobar 1.1. For any artist with an ampersand (&) in their name, the Last.fm context menu displays the first item ("artist top tracks") normally, but all other items have the previous item prepended, making for some crazy long strings towards the bottom of the list. I can't imagine my foobar being a minor version out of date is the problem. Can anyone else confirm this?

foo_softplaylists

Reply #333
nope, that isn't happening here. (latest component / foobar)

foo_softplaylists

Reply #334
1. thanks 4 the pluging but

2. foo_scrobblecharts was far much faster (why is it so fucking slow?? around 5 minutes to find 20x similar artists to ac/dc and 50 tracks each )

3. i miss the abillity to overwrite the "similar artists" playlist, why to create a new one all the time -> sucks

4. the playlist which is generated is randomized... i have to sort by artist first. Foo_scrobblecharts did that better. how to change this?

i use following values:



any tips 4 me please


foo_softplaylists

Reply #336
but foo_scrobblecharts uses last.fm too, or? with foobar 0.95 ... try it out. Its far much faster...

edit: Did some Benchmarks:

Same database: Result (up to 50 Tracks / Artist -> 20 similar Artists)
~150000 mp3s well sorted

similar Artists charts playlist to "Rammstein"

0.95 + foo_scrobblecharts

result: well orderd playlist -> including Rammstein again -> 704 items

1. run -> 68 seconds
2. run ->  2 seconds

1.13 + foo_softplaylist

result: new randomized playlist -> Rammstein not included -> 651 items

1. run -> 340 seconds !! 
2. run -> 69 seconds

Results can be compared...

scrobbelcharts TXT
http://dl.dropbox.com/u/21136636/Result%20...obblecharts.txt

softplaylist TXT
http://dl.dropbox.com/u/21136636/Result%20...oftplaylist.txt


=> interesting, isnt it?  o.O

Summary:

1. scrobblecharts is 5 times faster
2. track result is better (ordered playlist, artist included again)
3. playlist result is better, i dont have to delete the result playlists all the time! sucks
4. quality of results is the same!


Edit2:

i need a portation of scrobblecharts to foobar 1.1.3 please


foo_softplaylists

Reply #338
can anybody tell me, why those plugins take less time in the second run?
isn't the list of artists/songs sucked down from last.fm calculated freshly on each run?

just curious...
0.95 + foo_scrobblecharts
1. run -> 68 seconds
2. run ->  2 seconds

1.13 + foo_softplaylist

1. run -> 340 seconds !!
2. run -> 69 seconds


foo_softplaylists

Reply #339
i wonderd too... maybe it saves the result in a temprary file but i can restart foobar and its feels like the second run again (only checkt on scrobblecharts component). Just strange.

Can somebody make this benchmark too?

u can get my empty portable foobar2k archive here

http://dl.dropbox.com/u/21136636/foobar2000.rar

edit:

would be nice if somebody of the softplaylist team would look into the source code. Beside the optimization of speed the following points are in my interest:

1. Result playlist sorted by artist and charts
2. Result playlist includes the artist again
3. Result overwrites the last "similar charts playlist"

maybe optional done by checkbox in the softplaylist settings. Should increase the usability

foo_softplaylists

Reply #340
I'd love for the source to be made available, there are a couple of things I'd like to fiddle with, and I could implement some of the simpler and more sensible requests too, to show my appreciation and support. I understand if you don't want to show us your code though, I know there's plenty of code I've written that I would hate for people to see...

foo_softplaylists

Reply #341
Will look into it when I have time. But it probably has to do with http caching and the change to the internal foobar http functionality at some point.
I always thought Last.fm has become slower and didn't bother.

I can't be the matching algorithm because that is serveral order of magnitude faster than the one in foo_scrobblecharts.

It can also be that foo_scrobblecharts fetches less top tracks per artist and with less pages.

foo_softplaylists

Reply #342
thanks for the reply =)

I only compared the result for "Rammstein" - its similar as u can see by the results i added to my post. I would be happy if you can finde some time to look into the source.

I think the database algorythm is faster in the new foobar, the search is slightly faster. I cant say where the problem is. Maybe its possible to learn out of the source of scrooblecharts , i added it before too. Maybe its possible to copy the similar artists function nearly 1:1. Sadly foobar crashes with the usage of scrobblecharts but might be a little problem because the search works fine so far and if i ignore the failture in VS foobar2000 starts playing the new list normally.

Maybe anyone can portate it to 1.13 and check the reason for this crash by debuging.

Thanks!


foo_softplaylists

Reply #344
1. hm take my foobar2000 0.95 from above, or make a new 0.95 portable install and include scrobblecharts component... -> http://dl.dropbox.com/u/21136636/foobar2000.rar <-

2. same with an 1.1.3 portable + foo_softplaylist

3. initialize the same music data...

4. Then choose an artist like "AC/DC"

5. rightclick on a song "last.fm" -> "similar artists charts playlist" or something like that... simmilar in both versions

6. measure the time

7. thx


foo_softplaylists

Reply #345
nope, that isn't happening here. (latest component / foobar)


I've just updated to 1.1.2 (the download link on foobar's home page points to 1.1.2, not 1.1.3!) and tried again.

I've narrowed it down; it only happens when right clicking something in the facets panel (foo_facets) of the default UI. Right clicking files in the playlist and elsewhere works as expected, just anything in the facets panel acts strangely.

Can anyone confirm this more specific reproduction?

foo_softplaylists

Reply #346
can anybody tell me, why those plugins take less time in the second run?
isn't the list of artists/songs sucked down from last.fm calculated freshly on each run?

just curious...
0.95 + foo_scrobblecharts
1. run -> 68 seconds
2. run ->  2 seconds

1.13 + foo_softplaylist

1. run -> 340 seconds !!
2. run -> 69 seconds


The second run is way faster because of http caching. I can't speak for foo_softplaylist, but foo_scrobblecharts uses the windows (that's IE) http functionality, including it's Caching. Open IE and clear the cache - it should be slow again .

Quote
Will look into it when I have time. But it probably has to do with http caching and the change to the internal foobar http functionality at some point.
I always thought Last.fm has become slower and didn't bother.

I'm not using the foobar http functionality. I'm calling the Windows API myself. I'm not saying you should do that (you should not ^^), but that may be the source of the poblem.

Quote
I can't be the matching algorithm because that is serveral order of magnitude faster than the one in foo_scrobblecharts.

Not going to lecture you on optimization, but: you're doing in wrong . Orders of magnitude still don't matter if you are sorting a list of <1000 entries once on a modern cpu. One tends to optimize every line of code, but you should only optimize the parts of your code that actually take time. and in plugin like this one, this is obviously the networking part. I have no Idea why my plugin should be faster, but maybe you will find the reason when you compare the source codes.
Otherwise I would suggest you do mulitple http requests at once (try not to break the http specs too badly here ). foo_scrobblecharts isn't doing that, but it will speed up the process a lot anyways (it also would have done in foo_scrobblecharts, obviously).


Oh and as no one else wants to say it: Krautmaster, please stop trolling (yes, you're trolling).
Quote
2. track result is better (ordered playlist, artist included again)

That's not better, but a design decision. If you don't like the design decisions of a component creator, ask for an option to change it - once! If he doesn't want to change it and you can't live with it, write your own component and stop trolling.
Quote
3. playlist result is better, i dont have to delete the result playlists all the time! sucks

Also a design design - see above.

foo_softplaylists

Reply #347
I would suggest you do mulitple http requests at once (try not to break the http specs too badly here )

I always asked myself what happens, when we overload the last.fm api (or website?) with "Similar Tracks Recursion Limit >5"
Of course, it will be damn slow and take about 30 minutes to fill a list of maybe more and better hits, but would'nt that be too much "stress" for the last.fm site?

i myself didn't have the guts to try it (still have set to 1 and tried 2 a short while), because i want to live long with that great pluging and don't course any trouble last.fm maybe think: oh, they're abusing us, we will ban that plugin from getting data from us.

foo_softplaylists

Reply #348
Quote
I can't be the matching algorithm because that is serveral order of magnitude faster than the one in foo_scrobblecharts.

Not going to lecture you on optimization, but: you're doing in wrong . Orders of magnitude still don't matter if you are sorting a list of <1000 entries once on a modern cpu. One tends to optimize every line of code, but you should only optimize the parts of your code that actually take time. and in plugin like this one, this is obviously the networking part.
You are correct, but the XSPF part of the plugin doesn't use inet access. That is the main reason I sped it up with binary searches.

last.fm's speed depends a lot on when you access it. The times for fetching my +-500 tracks loved playlist:
Yesterday evening: 30s
Now: 5s

foo_softplaylists

Reply #349
i'd change the limit to 200. they allow it for recent tracks so i think it's fair game for loved tracks as well.