afriend wrote: > Ok, done. I'll put this on hold until you've found the time to do your > tests.
Musly needs to be initliased with a random selection of about 1000 'style' tracks - which, I think, lets it know the style of music you have. I thought rather then just random, I'd pick 1 random track (in the range 90s to 7mins) from each album, and then other random tracks to make 1000 tracks. (This 1000 can be less if you have less tracks, and can be configurd in config.json). But my code did not take into account there being no albums in the SQLite db (as meta-data has not be saved yet). So, after the analysis it is (I think) using all your tracks as 'style' tracks - and this is why it was taking so long. Until I have tested my updates what you can do is: - Backup musly.db - Run the analysis script to update the metadata data only - ./musly-server.py --analyse m --meta-only - After that, just run the anlysis as normal - ./musly-server.py --analyse m Setting the 'style' tracks takes roughly 15 mins for me with ~20k tracks on an i7 laptop with SSD. Not sure what your hardware is, or number of tracks, but it should not take hours. afriend wrote: > Would be nice if the database file I have so far (almost all tracks = > analyzed) could be reused with the new/revised code. You should still be able to use the old DB. *Material debug:* 1. Launch via http: //SERVER:9000/material/?debug=json (Use http: //SERVER:9000/material/?debug=json,cometd to also see update messages, e.g. play queue) 2. Open browser's developer tools 3. Open console tab in developer tools 4. REQ/RESP messages sent to/from LMS will be logged here. ------------------------------------------------------------------------ cpd73's Profile: http://forums.slimdevices.com/member.php?userid=66686 View this thread: http://forums.slimdevices.com/showthread.php?t=112713 _______________________________________________ plugins mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/plugins
