hendrik

I understand that nodes pass the path, but how does the image data move between them? I guess you use the Mistika engine backend for conversions? So basically graph builds an internal project that gets rendered when queue is executed? Because decoder has to push decoded image data somewhere and encoder must pick it up, thus the part about intercepting the data flow. This question is partly motivated by idea of adding reformatting to flow, currently I think there is no way to do it, what goes in, comes out resolution, fps etc wise..?

Except that from for example QtGui I get only ~20 entries in its dir, but in reality QtGui module includes hundreds of items. So I'd say almost all of them are missing: https://www.riverbankcomputing.com/static/Docs/PyQt4/qtgui.html

Some things I have noticed in script editor: when autocomplete offers completion string and I accidentally type a lowercase character (that is the same as in string with different case), autocomplete will concatenate whole string, not just the part needed. For example, typing PythonQtd will offer PythonQtDebugAPI, but pressing enter results in concatenated wrong string: PythonQtdPythonQtDebugAPI for convenience: commands like print etc could use different, lighter color schema for syntax highlight, they are very hard to read on dark gray background; for convenience: when typing a command and autocomplete offers something, pressing tab could also apply the completion, currently only enter works. Regarding Qt access, it seems to me that not everything in QtCore and QtGui libraries are accessible from script editor, only a small subseet is presented. I have looked for PythonQt (which I believe is used for adding python support here), it is not accessible at all. Qt, QtCore, QtGui modules seem to only present a few select classes. Is there a reason for this?

Not directly python questions, rather about overall architecture of workflows. What kind of data is passed from node to node? Lets say I have a ExrFile input connected to prores process node. One of them is decoder, another encoder. So how exactly does image data move between them? And is there way to intercept it (the actual image data) with Python? What is the order or operations for executing nodes when processing? Lets say I want to modify the url of some input node with python. How can I accomplish this? I can't add python node in front of input node and placing it separately or after it does not execute it before...

EXRClip input does not populate "Frames" variable although it shows the frame range properly in url. It would be logical if frame count is updated to reflect the url. None of the other values besides resolution, aspect, fps and format are filled either.

Are there methods to read metadata from actual file, lets say I have a Prores or exr input, can I somehow access all the meta in file?