Parisson / TimeSide

scalable audio processing framework and server written in Python
https://timeside.ircam.fr/docs/
GNU Affero General Public License v3.0
369 stars 59 forks source link

Problems installing in osx ~ "gobject missing" #45

Closed vectorsize closed 9 years ago

vectorsize commented 10 years ago

I've been trying to install timeside on my osx machine and I made it pretty far (after cumbersomely installing several dependencies). Ive reached to a point where I can include timeside inside a python script, except now it complains about gobject missing.

Here is the thrown error:

import timeside
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "timeside/__init__.py", line 4, in <module>
    import core
  File "timeside/core.py", line 31, in <module>
    import gobject
ImportError: No module named gobject

I tried brew's libgobject, pygobject and gobject-introspection without any luck… any ideas?

vectorsize commented 10 years ago

brew install pygobject --with-python didn't help either

yomguy commented 10 years ago

Hi Victor!

Installing TimeSide from scratch on OSX is indeed a pain. Sorry for that, we all run it over Linux here. Maybe @piem could have something to say about this (last ?) dependency pb for gobject.

Are you running python 2.6 or 2.7?

Before we get a nice install procedure on OSX, maybe @ouhouhsami could install timeside on one a your Linux server so that you can run it through ssh (as you would do in a local virtualbox) ? It would take indeed a few seconds to install following the README ;)

vectorsize commented 10 years ago

I'm running 2.7.2.

I was hoping to help with finding a "nice install procedure on OSX" here :) As far as doing a linux install, I can take care of it, don't worry, now I just want to find the way on osx!

yomguy commented 10 years ago

2.7.2 is OK.

It would be great indeed if you could keep your bash_history so that we can finally add the howto to the doc. Thanks a lot for that Victor (and for your patience) !

yomguy commented 10 years ago

Maybe a PYTHONPATH tuning? http://www.linuxquestions.org/questions/linux-software-2/importerror-no-module-named-gobject-exaile-fails-to-run-python-problem-569303/

yomguy commented 10 years ago

Can you then check the gobject module is in your $PYTHONPATH ? On Debian it appears in : /usr/lib/python2.7/dist-packages/ for example..

yomguy commented 10 years ago

As @ouhouhsami proposed:

brew install pygobject --with-python

https://github.com/Homebrew/homebrew/wiki/Homebrew-and-Python

vectorsize commented 10 years ago

didnt work :( …still in process, now it's pygstmissing

vectorsize commented 10 years ago

also python path wasnt set, now it is

yomguy commented 10 years ago

ok, try:

git clone git://anongit.freedesktop.org/gstreamer/gst-python
cd gst-python
./autogen.sh
make
sudo make install

?

piem commented 10 years ago

I was just updating the formula for aubio...

let me have a look!

On 14 Mar 2014, at 09:43, Guillaume Pellerin notifications@github.com wrote:

ok, try:

git clone git://anongit.freedesktop.org/gstreamer/gst-python cd gst-python ./autogen.sh make sudo make install ?

— Reply to this email directly or view it on GitHub.

vectorsize commented 10 years ago

brew tap homebrew/versions brew install gst-python010 made it for pygst

now dealing with other problems :) will keep you posted

piem commented 10 years ago

@vectorsize try this formula to build aubio 0.4.1: https://github.com/piem/homebrew/tree/master/Library/Formula/aubio.rb

i'm waiting for the pull request to go be accepted: https://github.com/Homebrew/homebrew/pull/27509

next step will be to build yaafe, and you should be set.

piem commented 10 years ago

sorry, i meant to point at the aubio branch:

https://github.com/piem/homebrew/tree/aubio/Library/Formula/aubio.rb

vectorsize commented 10 years ago

thank! yes found the raw path, the formula works, however I'm not sure that's the problem?

piem commented 10 years ago

@vectorsize where are you at now? are you still missing gobject?

vectorsize commented 10 years ago

it seems to have worked but then when I import timeside h5py complains:

Traceback (most recent call last): File "<stdin>", line 1, in <module> File "/usr/local/Cellar/python/2.7.6/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/h5py/__init__.py", line 10, in <module> from h5py import _errors ImportError: dlopen(/usr/local/Cellar/python/2.7.6/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/h5py/_errors.so, 2): Symbol not found: _H5E_ALREADYEXISTS_g Referenced from: /usr/local/Cellar/python/2.7.6/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/h5py/_errors.so Expected in: flat namespace in /usr/local/Cellar/python/2.7.6/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/h5py/_errors.so

yomguy commented 10 years ago

Hi @vectorsize, Sorry, we are late on this bug, but we lack some Macs here to test the OSX install.. Maybe some clues here for this problem about h5py: http://stackoverflow.com/questions/9873427/h5py-gives-error-after-installation ?

yomguy commented 10 years ago

Hi @vectorsize, Would you be ok to run our Vagrant box or Docker image? https://github.com/yomguy/TimeSide/tree/dev#virtualbox-and-vagrant https://github.com/yomguy/TimeSide/tree/dev#docker

yomguy commented 10 years ago

On Mac, the best way now it to use our Vagrant box or Docker image.

gzepr commented 9 years ago

virtualization is a workaround, not a solution. @vectorsize, did you manage to get through ?

yomguy commented 9 years ago

yes @gzepr but a workaround is even better than a fail ;)

gzepr commented 9 years ago

true @yomguy but thanks for reopening, timeside looks great and i would love to use it on my mac

vectorsize commented 9 years ago

sorry @gzepr I didn't manage, haven't had time to try again…

yomguy commented 9 years ago

This will be finally fixed with the docker image through the 0.8 release.

SamridhiJ commented 6 years ago

how to fix this in windows ... please help me out ..

yomguy commented 6 years ago

Hi @SamridhiJ, please follow the new setup process thanks to docker-compose and not the old one.