found a simple script that did not read all tags. It only use the filenames in artist - album - song structure. This script full rescan 9000 files in under 15 minutes. The right folder structure is coming from the banshee player, it can move and rename structures by tag, Mediatomb can rescan such changes very fast.
Someone out with such script that will sort artists to abc groups like described at the help pages? I'm really a starter with the script language and dont want tag reading.
You mean an import script for MediaTomb?
The speed does not come from is *not using the tags*, it comes from *creating less virtual objects* in the database.
When the scripting engine gets an item, the item already is populated with all metadata that we could find. However, it's the number of virtual items that are created by the script that take so long, i.e., for one imported audio item the script can create about 6 virtual items in various containers (i.e. by artist, album, genre, year, etc…)
The import2.js from boilerjt creates only an audio chain with abc - artist - album - song. Took around 20 minutes with 10k files. Maybe a simpler chain fixes several problems. With reciva radios you run in "browse request failure" when a very long list (1000 artists) are reloaded while feeding a playlist. The radio fails while playing well from the server. This can be a timeout or buffer overflow, was told to reciva. Simple chains with abc diversion give easy and fast methods to fill a playlist in player devices. Idea: Bundle different import scripts to the package.
> Idea: Bundle different import scripts to the package.
Nothing against that, actually that's what the scripting Wiki is for, but of course if users submit full and complete scripts we could add them to the package.
Forgive my ignorance but where can one find this "import.js from boilerjt"? I have the same issue you do, I don't need all of the extra meta info being checked (or virtual objects), I just navigate using the PC directory function. Right now when I add music it takes days for it to catalog just a few albums. I'm running mediatomb on a nas so with the limited cpu this means I can't stream anything properly while it is cataloging.
Use this archives and follow the instruction to install.
This version comes with scripting support to use own variations, the original ffp mediatomb is compiled without scripting and cannot solve this. I have 12000 files and 2000 pics now and the simpler scripts read all in 20 minutes. But it's a static build, very huge executable that eats nearly all mem on the nas. While scanning the usage is sluggish. I'm on firmware 1.08 with the 323 now, the mt-daap and firefly are crap like before, only mediatomb give a proper environment to my reciva radios.
I'm using this on my NAS (aka the dealextreme nas or wlx-652) so I'm not sure if this build would work. The version I'm currently running was downloaded from here http://downloads.sourceforge.net/mediatomb/mediatomb-static-0.11.0-r1-linux-uclibc-armv4.tar.gz
I've looked and I do have an import.js file, would I be able to edit this or replace it with one another user created to achieve what I'm after?
The script interpreter seems to be the same in all builds, so it must use every available script. I use the boilerjt script with some modifications, to get a simpler navigation on my recivas. Hope your build is with interpreter activated?
Is there an easy way to tell if my build has the interpreter activated?
I found this little snippet in my config file, I thought I'd paste it just in case it is helpful. I was thinking that maybe the "builtin" part might indicate that it can only use "built in" scripts but hopefully I'm wrong.
Guys, don't guess - read the docs!! And the docs say: change "builtin" to "js" if you want js scripting to be used. If js is not available while you setup the config to use it - it will complain anyway.
Log in to post a comment.