Am Montag, den 01.08.2005, 18:53 +0200 schrieb Joerg Lehmann:
Hi Doug,
On 01.08.05, Doug Friend wrote:
Thought I'd introduce myself. Just installed PyTone 2.2.4 & have it running after a couple of tries. On initial set up I pointed it to 120GB of music which caused some fairly severe performance issues. Anyone have experience with PyTone & a large music collection?
I've heard of performance problems with huge databases. Maybe optimizing the cache settings helps. Have you tried to increase the cachesize of the bsddb (option "cachesize" in the corresponding database section of the config file). If you observe a performance degradation after a certain time, it may be worth to decrease the database request cache (currently, this is hardcoded in .../services/songdb.py, line 115: self.resultcachesize). I plan to move this to make this configurable via the config file and also to add some statistical output on the cache usage).
I'm new to Python, have been coding with Perl for the last 5 years or so. I have been planning to write a playlist generator which will take additional parameters than what is currently in PyTone (tempo, texture, music style among others). Just starting to get into the code now, so any words of wisdom would of course be more than welcome :)
As Richard has pointed out, the set of tags stored by PyTone is currently limited to the usual track/album/artist/year/genre. Adding more metadata should not be too difficult, however.
I'm wondering if the database layer could be written to support for example ZODB. This could improve some things, for example we could avoid trashing the database on powercycle as bsddb likes to do unfortunately. Relatinal dbs would be nice too but a lot more work I suspect. Greets Tino