Ahh cortex, i have done work with that myself in the past. And i am sure we
used cortex bindings to write the necessary data directly in python into a
file. Again, camera information is all accessible form python you need the
matrix and the projection attributes and they are all ready for you to use
in python. But yeah if you want that function form cortex i would suggest
you write a node that can take a camera input and then cast the input to a
camera Op check the pointer call ->validate( true ) and then pass the
pointer into your cortex converter. and all you do in python is execute
that node. This does not need to be anything else but executable. ( this is
a class you can derive from in the NDK ) The class is DD::Image::Executable
and you should find good examples looking for that class in the examples
and the NDK docs. This class gives you ways to execute that node like a
write node and then call a function per frame to do the exporting. This
would be the cleanest solution even if you just create that node in python
in your exporter as a temp node that you delete after exporting.

RE: your request for an API for the NODE itself. This is done on purpose
like this as far as i know. You have python to deal with nodes. It could be
rather dangerous if you had C++ access to those nodes. I do agree that
reconnecting nodes could be done easier then calling python from c++ but as
i have now  a little lib of functions to do exactly that ( get node by
name, get param by name, connect node, disconnect node ) i am happy with
it. I might be able to share this soon  here so we have it written down.
Its all super simple but i agree it would be nice to just have it somewhere
ready to use.

Take care
jo

On Sat, Mar 7, 2015 at 9:24 AM, Lucien Fostier <[email protected]>
wrote:

> I'll check if there is an API for sceneCache camera to be created from
> intrisic values.
>
_______________________________________________
Nuke-dev mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev

Reply via email to