Hi Roland. I've been playing with creating a Don't Stop The Music mixer using Musly. For this I've create a python script to analyse files with musly and provide a 'get similar tracks' API via Flask. I've used your musly.py code to interface with musly, but added usage of python's concurrent features to speed up analysis. On my 5 year old i7 laptop with an SSD card I can analyse ~20k in just over 1 hour - which is amazing compared to the ~1week for MusiIP (although that was not 24*7). Looking at the timing of the analysis, the last ~20minutes are taken by the call to setmusicstyle. This is not that bad when analysing all tracks, but setmusicstyle is called if more tracks are added - and takes ~20minutes again. Do you know of anyway to speed this call up? I tried re-compiling Musly with OpenMP but this made no difference.
I case you are interested, the code for the API server and DSTM mixer are here: https://github.com/CDrummond/musly-server https://github.com/CDrummond/lms-muslymixer ...please be aware that I have not extensively tested this, and they're probably no where near ready for prime time - but do seem to work OK in my limited testing. *Material debug:* 1. Launch via http: //SERVER:9000/material/?debug=json 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=108495 _______________________________________________ plugins mailing list plugins@lists.slimdevices.com http://lists.slimdevices.com/mailman/listinfo/plugins
