Hi,

Sorry if this is a silly question, but in all my Googling and looking through the docs, I couldn't find anything related to this besides a vague mention in the GIMP-Python documentation.

I'm writing a plugin that scales an image to 2x size using Hyllian's Super-xBR algorithm.

I have an array of RGBA values that I want to draw into a pixel region that spans the image. Formatted like this:


where indices 0-3 are the R, G, B, A values of the pixel at [0, 0], indices 4-7 are the values of [1, 0], etc.

The documentation says that the input data for setting a pixel region "is a string containing the binary data of the requested region", but I'm not sure how it should be formatted. The plugin I used as a basic reference, Akkana's arclayer.py (can find on their website at shallowsky), just uses array.tostring() to change the input into a string and then sets the pixel region equal to it:



however, doing this simply gives me an empty(transperant, all RGBA values 0) layer when I insert the layer into the image.

Is there another, or better, way of doing something like this? Or do I just have the wrong idea completely?

This is my first time using Python or dealing with GIMP at all(I used paint NET pretty heavily before I switched off Windows), so apologies if I'm missing something obvious.

Here's some relevant code, if it helps:



Thanks.

My experimental numpyops.py script reads pixel regions into numpy arrays. Source is here: https://www.gimp-forum.net/Thread-Gimp-python-and-numpy

Thanks for the source/tip, I noticed a lot of people seemed to recommend numPy arrays instead of using python's stdlib array class.

I'd like to keep it as portable as possible, though, without relying on external dependencies. Although Super-xBR is quite slow in Python, so if numPy was the only option, I wouldn't have too big a problem with using it.

Apologies for the double post, but I've been continuing to look at this and it seems that the output data is actually copying into the pixel region array just fine, as it should. However, no matter what PDB or Image methods I call(I've tried update and flush), I've been unable to make the pixels write to the layer - it continues to simply insert a blank layer and never fill the layer with the pixels in the pixel region, despite the layer having a reference to the pixel region.

Any help would be appreciated.

The first script is a template I made for using region derived from Akkana's arclayer.py. From this template I made the second script, I don't know if it will help with your problem but you might get some ideas from it

The Region test script swaps the red and blue channels, done just to see it works

These scripts work on my Gimp 2.8.22



scripts are found at "File > On test >"

[EDIT] I should have mentioned that in your coding you use 4 spaces which is convention, being lazy I use tabs. The thing is you can't mix them, needed to state this just in case you copy/paste any lines of my script. If you do, change each leading tab to 4 spaces

steve

Could be a simple error in your code, but until you post it we can't tell.

From my own script:



Which more or less is summarized as "get a pixel region from the layer and copy the bytes to it".

Your reference code was helpful, thanks. I was under the impression that merge_shadow() wasn't important, but that was because I wasn't aware that the 'shadow' refers to a temporary processing buffer, and that without merging it, the layer will not be updated. After adding a call to merge_shadow(), the layer is successfully filled.

Guess I should have looked it up in the procedure database.