Jump to content

Lorenzo Angeli

Administrators
  • Posts

    331
  • Joined

  • Last visited

  • Days Won

    32

Everything posted by Lorenzo Angeli

  1. @jen_at_floyd we are looking on adding your proposed fix (with few changes) to the api 2.0 branch. If you want you can test this before cloning the repo , changing branch and installing it into a virtualenv. This fix is likely to end up in the api 2.0 branch soon though, let us know in case you test. Cheers. L.
  2. Glad to hear @Konstantin Maslyuk ! The branch still requires some tweaks and details, but should soon get to master. More news and info on the branch itself L.
  3. @Konstantin Maslyuk I've now made so calling build_ext use the same command as Build hence building the missing resource.py, hope this ease your problem. This is what I'm doing for testing create a new virtualenv (I personally use mkvirtualenv wrapper) activate virtualenv go with the shell in the ftrack-connect repo folder and checkout the given branch (and pull) run to install in editable mode : pip install -e . run ftrack-connect : ftrack-connect here the versions of the basic moving parts I'm using: python : 2.7.14 setuptools : 44.0.0 pip: 19.3 (20 works too) Let us know how it goes. L. p.s should work in docker too, could you share your dockerfile so I can check ?
  4. Hi @Konstantin Maslyuk I've just been pushing a new commit on the above PR to ensure you can install qtext in build_ext command. Please pull again and give it a go. let us know how it goes. L.
  5. Hi @jen_at_floyd, thanks a lot for raising this issue and providing a potential solution. We'll be looking on whether this is something we can push upstream on the final release ! Cheers. L.
  6. Hi @Konstantin Maslyuk although not tested, connect should be able to support pyside2 though our QtExt layer, but has not been reviewed for supporting Python 3k atm. I'll be double checking the build_ext functionality in the above branch, keep monitoring it. > The rest is fine, i used docker/ubuntu:18.04. connect is currently tested and built for centos 7 , I've not much of an experience on Ubuntu, so not sure what packages are available and to what version. for this error: > distutils.errors.DistutilsError: Could not find suitable distribution for Requirement.parse('qtext') please try to downgrade your pip to version 19.3 and try again. Let us know how it goes. Cheers. L.
  7. Hi @Konstantin Maslyuk, this is something I need to double check as I usually go for using pip to install dev versions with : pip install -e <path to repo> regarding connect there's a branch I'm looking into these days similar to your request: https://bitbucket.org/ftrack/ftrack-connect/branch/backlog/review-standalone-install try to give it a go and let me know. Cheers. L. p.s Jinja 3 is not a requirement of ours, is likely coming as sub dependency of sphinx or something similar. if you have a chance to pin down the offending module version let me know.
  8. New RC version is out for testing.... this version introduce a better check for remote event server connection while listening to events. you can install the new version with : pip install ftrack-python-api==2.0.0rc3 We are getting close to a final release, please report any issue you may find during the RC release phase !
  9. Thanks to all your feedback and bug reports, we are finally getting closer to a final release of the upcoming api 2.0 for python2/3 . If you want to install (at your own risk) the latest RC2 version please use the following command: pip install ftrack-python-api==2.0.0rc2 Full change log can be found here please report through the usual channels any issue or bug you might be finding. Cheers. L.
  10. I'd suggest to stick to C7 as long as possible atm. By direct experience C8 is not ready yet for production. L.
  11. @Nero Demaerschalk, Centos 8 is something we'll be looking into as soon as the distro is stable for desktop use (I'd suggest waiting more than .1 release) For python3 we have a dedicated thread in the forum at this address, feel free to test and let us know if you find any issue. We are working in these days to release a new api 2.0 for python 2 and python 3 for wider use, but the inclusion in connect will require more time due to some non backward compatible changes in the api. Hope it helps. L.
  12. Hi @jen_at_floyd, there's a patch already merged in connect to mitigate this and it'll be available in an upcoming version of connect.
  13. Hi @Tilt , having the python version to >= 2.7.9 has always been a soft requirement for the api, but was never enforced through the installation process itself. Once the amount of issues raised by having python not correct has got to an alarming level we decided to raise the requirement to hard. My suggestion would be to install a newer version of python on C7 (quite a straight forward process). Hope it helps. L.
  14. Hi @Chandler there no boilerplate at the moment, but I'll try to find time new year to come up with one. As you find out though, there was one once, and is the one I'll try to bring back to life or close to what that was at least. Hope it helps. L.
  15. Hi @Alican you cannot use the location.get_url(component) as path for the component as is relative to the server, path requires to be a local path to your file system. What you might be after, if I'm getting it right , is an event which collect the data which might have been manually uploaded and then transfer them first to you local disk , before being re uploaded as reviewable. For this to work , you need to have a listener running which will intercept the update event for components and use that to move the data from the ftrack.server to the local storage. Something along these lines should work (https://bitbucket.org/snippets/ftrack/ynxGp4) Hope it helps. L.
  16. HI @Alican just gave it a go with default connect and actions and I can see them appearing on both Chrome instances I've open (application launchers). May I ask where the actions are hosted ? On each ftrack-connect-package plugin folder (local to the user) , in a shared place and mapped through FTRACK_CONNECT_PLUGIN_PATH / FTRACK_EVENT_PLUGIN_PATH or are running as standalone event listeners ? Also, to help us investigate, removing custom events and actions does still trigger your issue ? Looking forward to hear more from you . L.
  17. HI @Dimitri Lucy, glad to hear ! let us know if anything else comes up L.
  18. Hi @Dimitri Lucy, whenever you are manipulating the PYTHONPATH I suggest extending it rather than overwriting it with something like : PYTHONPATH=%PYTONPATH%:C:\Python27\Lib\site-packages\ftrack_api On the error you are having though, seems to me that you are installing the new api , but trying to import the legacy ones. Could you please try to import ftrack_api rather than ftrack ? Let us know how it goes ! Cheers. L.
  19. HI @Mani Varma, If all you need is an efficient api to represent a tree I'd suggest having a look at this widget from connect. It does use a qtmodel which is quite efficient. Hope it helps. L.
  20. HI @Milan Kolar, Thanks for reporting it. I wasn't aware you were already using it ! Be aware that some breaking changes (session won't connect automatically to the event hub) are coming soon on the story branch so be careful. Feel free to get in touch through support to share any stack trace you might be getting. Cheers. L.
  21. Hi @Julian Martinz looks like we are on the same page now let us know how it goes! L.
  22. Hi @Julian Martinz, you cannot create two locations with the same name that would confuse the system. The idea of the proxy is exactly for this reason, the proxy will have your old name to match (studio.server1), where the underlying (proxied) locations will retain the different behaviour (studio.server1.managed, studio.server1.unmanaged) and the logic in the overwritten methods of the proxy location will dictate what component to handle with what location. Hope it helps. L.
  23. We have been restarting the work on the Python api with Python3k support. There's a new branch ( backlog/api-2.0/story ) to use to pick the latest and greatest, and it'll be used until the official release. Here how to update the above snippets to use it: ./pip3 install git+https://bitbucket.org/ftrack/ftrack-python-api.git@backlog/api-2.0/story If you do have any question of you find any issue please let us know.
  24. >> What would work is creating two distinct locations studio.server1 and studio.server1.unmanaged. Hi Julian, my suggestion of having a third location acting as proxy was coming from the idea of leveraging two different location with different behaviour. This though , requires to have clear rules on how these two locations will accept data (eg: image sequences -> unamanged, maya files-> managed.) >> Is there any way to revert a location instance back to the state it was before mixin in the unmangedLocationMixin? Not that I'm aware of tbh.... the above solution was though to be able to provide both at the same time. >> reconfigure my locations by emitting an 'ftrack.api.session.configure-location' event. This is something should be called when initialising the session not sure doing it programmatically would provide the results you are after. L.
  25. Hi @Jakub Trllo, you have almost answered yourself the question: >> It's probably because every time is used `custom_attributes` key on entity, api is querying values of custom attributes? It is quite possible that is what causing your slowdown. If you don't need / care about what the previous value was you can think of turning off the auto_popluate using the context manager for it. You can find further information here. Let us know how it goes! L.
×
×
  • Create New...