If you take a look at the source for selectConnectedNodes() (in
nuke.overrides), you'll see that all it's doing is walking the DAG. :) For
various reasons (efficiency, recursion, more functionality, generators), I
ended up writing my own traversal functions. There really isn't another way
to get a node's dependencies unless every node were keeping track of all of
its dependencies at all times... which would completely defeat the purpose
of building a DAG in the first place.
One thing to keep in mind is that selectConnectedNodes only operates on a
single node, so to get accurate results, you'll need to call it once for
each Write; you may be getting the results you're expecting by accident
right now if all of your Writes branch from the same tree. Since it looks
like you don't care care about isolating the dependencies for each leaf
(Write) node, the code I posted can easily be adapted to find the common
dependencies for an arbitrary set of initial leaf nodes (just populate the
'stack' with the union of all of their dependencies before starting the
traversal).
Hope this helps.
-Nathan
-----Original Message-----
From: Jep Hill
Sent: Thursday, January 30, 2014 8:39 PM
To: [email protected]
Subject: [Nuke-python] [nuke -t] oddness when opening
scriptswitherrorsininteractive mode
Hey Nathan -- thanks for that perspective... I was thinking that scanning
for writes that aren't disabled, then using the selectConnectedNodes method
proves to be a very fast approach for isolating the nodes in play. Once
that's done we don't necessarily need to walk the DAG since we already know
the pertinent nodes, right? Do you think it's possible to leverage the speed
of already knowing the nodes in play to accelerate the process even more?
let me know if you think we should roll this over to another thread...
Cheers,
Jep_______________________________________________
Nuke-python mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-python
_______________________________________________
Nuke-python mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-python
