Lorenzo Angeli

Answered in support case, but here for reference: For what concerns location accessors, is pretty doable to write S3 and Azure ones (it has already been done before), you can find further information looking into the api: https://bitbucket.org/ftrack/ftrack-python-api/src/master/source/ftrack_api/accessor/ If the cloud storages provide a python api access (eg: boto for Amazon S3) and using the base accessor ( https://bitbucket.org/ftrack/ftrack-python-api/src/master/source/ftrack_api/accessor/base.py) as baseclass , is just matter of fully re implement the various method exposed. Although I would suggest to use these as a mid ground storage and keep the production files locally to the office and use these other locations just for transfer, this would allow you to get the best of both worlds. Hope it helps. L.

Hi Reynaldo, answered in support case, but for reference: Although the official support for python 2.X is set to 2020, DCC company applications (Autodesk, Foundry etc...) will slowly move over. During last year's Siggraph we've been discussing with them and through the https://www.vfxplatform.com it has been decided to start putting ahead a 3.6.X but is not been nailed yet, and therefore it'll likely take way longer to get fully adopted by the industry (I'm expecting a delay of at least 2 years.) To be honest I would avoid moving to py3 before all the DCC apps have been doing it. If you are interested though, you can find our (not ready to merge, therefore not supported officially yet) branch for py3k compatibility here: https://bitbucket.org/ftrack/ftrack-python-api/pull-requests/191 Hope it helps. L.

Cheers ! Let me know if you have any other issue or question.

Hi Kristing ! I've been looking now at the example you posted, and looks like it's missing something in order to cope with what you are after. Here how I got it working. In: ~/.local/ftrack-connect-plugins/testactions/hook * first.py * second.py each of these files contains a copy of the linked example. Here is how the classes are initialised: class MyCustomAction(object): '''Custom action.''' label = 'My Action' identifier = 'my.custom.action' description = 'This is an example action' variant = 'first' class MyCustomAction(object): '''Custom action.''' label = 'My Action' identifier = 'my.custom.action' description = 'This is an example action' variant = 'second' Then change in (both) the discover function to include the variant : return { 'items': [{ 'label': self.label, 'description': self.description, 'actionIdentifier': self.identifier, 'variant': self.variant }] } rather than : return { 'items': [{ 'label': self.label, 'description': self.description, 'actionIdentifier': self.identifier, }] } If you then restart connect , this is how the action should show up: To simplify further though, I would also suggest you to have a look at : https://bitbucket.org/ftrack/ftrack-action-handler Which aims to simplify the implementation of custom actions (http://ftrack-action-handler.rtd.ftrack.com/en/stable/). Please let me know if you have any further issue or questions. Cheers. L.

Cheers Scott, do you mind linking the thread on thinkbox forum ? So I can read and, in case, follow up ? If you have some development knowledge, or you have developers available, you can think of customising their code.(ps. sadly cannot find it on the github repo : https://github.com/ThinkboxSoftware/Deadline) Thanks. l.

Hi Scott, Deadline integration is maintained by ThinkBox itself, and I've not much direct experience with it. Regarding your issue though, the options :Submit Render Layers As Separate Jobs is enabled. sounds fishy, as is likely to generate one job (and therefore a new asset) for each layer. To me the best option would be to have 1 asset (Render) with all the render layers / passes published as components of the given asset version so would easy to know where they come from. e.g: ``` City (Render) v010 - diffuse: <path to diffuse pass> specular: <path to specular pass> <other_render_layer_name>: <path to other pass> ``` and so on. I'll try to dig deeper as soon as I have a chance in this integration and try to get a better grip on how could be improved. Let me know how it goes please. Cheers. L.

Hi Robin ! Although totally doable , with an event that intercept the publishing event and update a folder with a symlink , or copying to the latest version, this is not something I would suggest doing. When you embrace an asset management system you should almost forget the file system structure and rely on the server queries* to see what's the latest/current version. This because latest or current, could not be the latest published but , possibly , the latest approved, or not even on the same disk. If you always have to rely on latest published, my suggestion would be to code a small tool to pick the latest published through the api only. Hope it helps. If you have any other question, please let us know ! L. * eg: the file system where assets are published is not available to the artists to look into, or based on id rather than names, so the published files cannot be modified.

Hi Tuan, Ftrack default (storage scenario / default structure) file system structure is sanitised in order to work in mixed operating system environment. here where does happen: https://bitbucket.org/ftrack/ftrack-python-api/src/c2fcbca6f8a957a423f6bc2f2bde768a0d8e50b6/source/ftrack_api/structure/standard.py#lines-126 The best solution you have to get this as you like, would be to create a custom location + structure. If you want to dig deeper on how to create custom location (and structure), have a look in here : http://ftrack-python-api.rtd.ftrack.com/en/stable/locations/overview.html?highlight=structure http://ftrack-python-api.rtd.ftrack.com/en/stable/locations/tutorial.html Hope it helps ! L.

Hi @hulmanen & @AnthonyM this is a good point. I'll make sure is going to be added somewhere sensible in the docs. Cheers. L. p.s Cheers @Remus Avram

Hi Lori, if you are using rez, you can have a look in here for ftrack-connect 1.1.3 package: http://git.efestolab.uk/rez/rez-ftrack/blob/master/ftrack_connect/1.1.3/package.py The web ui path is resolved through this hook contained in ftrack-connect resources in :https://bitbucket.org/ftrack/ftrack-connect/src/master/resource/hook/resolve.py Add that folder to FTRACK_EVENT_PLUGIN_PATH and PYTHONPATH and you should get your paths resolved. If you have any issue please let me know. Cheers. L.

Hi Anthony, can you please check you have libpng installed ? Once installed please try launchihng connect again and let me know. Cheers. L.