Plugin is not found in the menus

Krikor · **Joined:** Dec 19, 2018 **Posts:** 164

Hello,
I would like to test this plugin but it does not appear in any menu in GIMP.
Even searching via Help - Plug-in Browser - Search I cannot find the plugin.

Could any Python programmer analyze this code and check if the way it addresses the plugin location is correct?

Code below in spoiler.

Thanks in advance

SHOW SPOILER

MareroQ · **Posted:** Fri Feb 07, 2025 7:57 pm **(#2)**

You must publish as "code" or provide a source link. :hehe

Krikor · **Joined:** Dec 19, 2018 **Posts:** 164

MareroQ wrote:

You must publish as "code" or provide a source link. :hehe

Hello MareroQ!

I'm sorry but I don't understand what "publish as "code" or provide a source link" means. :roll:

I'm really a noob when it comes to plugin code.

Let me clarify the situation here.

Some friends and I were discussing which AIs (GPT, DeepSeek and Qwen) would be the best to use.

So we subjected them to the same tasks and compared the results.

Since I've always admired this effect in PS, I decided to test if these AIs would be able to produce Blend-If for GIMP. But I wasn't even able to find the plugin after installing it.

I even tried to tweak the code. I made some intuitive changes, reported the problem to the AI, but in the end I was still in the same situation as before.

It's possible that the code won't even work even if I manage to address it correctly so that it appears in the Filters menu.

In any case, I tried, I did what I could.

MareroQ, thank you very much for answering and contributing to my question! :bigthup

ofnuts · **Joined:** Oct 25, 2010 **Posts:** 4812

Python is sensitive to code indentation.

Code: Select all

if condition:
    print("this is printed if condition is true")
    print("this is also printed only if condition is true")
print("this is always printed")

The code you show is like this:

Code: Select all

if condition:
print("this printed if condition is true")
print("this also printed only if condition is true")
print("this is always printed")

So 1) python complains that it doesn't see an indent after the "if condition" line and 2) figuring out which line(s) should be re-indented requires some serious reverse-engineering/debugging.

So either the code you have has already lost its indentation and you are in trouble or it still has its indentation and then provide it in way that preserves it; either using \[code\] \[/code\] tags as I did above, or by adding the Python file as an attachment.

Krikor · **Joined:** Dec 19, 2018 **Posts:** 164

ofnuts wrote:

Python is sensitive to code indentation.

Code: Select all

if condition:
    print("this is printed if condition is true")
    print("this is also printed only if condition is true")
print("this is always printed")

The code you show is like this:

Code: Select all

if condition:
print("this printed if condition is true")
print("this also printed only if condition is true")
print("this is always printed")

So 1) python complains that it doesn't see an indent after the "if condition" line and 2) figuring out which line(s) should be re-indented requires some serious reverse-engineering/debugging.

So either the code you have has already lost its indentation and you are in trouble or it still has its indentation and then provide it in way that preserves it; either using \[code\] \[/code\] tags as I did above, or by adding the Python file as an attachment.

Hi Ofnuts,
I should have noticed that the code was missing indentation when I pasted it here in this post.

After your observation, I was wondering if the lack of indentation was already generated by the AI or if I lost it at some point between pressing ctrl+c and ctrl+v.

I checked and found that the AI generated the code correctly indented.

So, while I was typing this text, I decided to check if the code in the plugins folder was indented or not. Yes, it has the correct formatting!

I'm going to paste the code again and if during this process it loses its formatting (indentation), I'll attach the .py.

SHOW SPOILER

Thank you very much for your help!

PS.: Oh.. I noted the following when creating this post:
When I paste the code in the text area between s poil.../s poil, and then preview it, everything is correct, the indentation is present! But after submitting the post for publication somehow the indentation is lost.

Attachment:

Indentation02.png [ 158.13 KiB | Viewed 3987 times ]

teapot · **Joined:** Dec 09, 2018 **Posts:** 736

Don't use spoiler tags, use code like this:

Code: Select all

blah
    indented blah

The code icon is two to the left of the spoiler icon.
Or you can just type the tags.

ofnuts · **Joined:** Oct 25, 2010 **Posts:** 4812

These damn foreigners:

Code: Select all

python blend-If.py 
  File "blend-If.py", line 6
SyntaxError: Non-ASCII character '\xc3' in file blend-If.py on line 6, but no encoding declared; see http://python.org/dev/peps/pep-0263/ for details

Add this as the second line:

Code: Select all

# -*- coding: UTF-8 -*-

This said, looking at the code, 1) it should be very slow (can be improved somewhat using "pixels regions" and 2) I wonder how different this is from this:

* Copy the two layers
* Blend them (top layer at 50% opacity then merge down)
* Apply over the original bottom layer in lighten only mode.

More accurate procedure:

* Duplicate both layers and apply "Desaturate > Luminosity" to each
* Set the top of these to "Grain extract" and merge
* Threshold the result to 128 or .5, this gives you a map of the pixels where the top layer is more luminous than the bottom one.
* In the Channels, do a "Channel to selection" on any of the RGB layers
* Back to the original top layer, add a layer mask and initialize to selection
* Set the opacity of that layer to 50%

Can of course be scripted, but on any large image it could be faster manually than running the script.

Krikor · **Joined:** Dec 19, 2018 **Posts:** 164

ofnuts wrote:

These damn foreigners:

Code: Select all

python blend-If.py 
  File "blend-If.py", line 6
SyntaxError: Non-ASCII character '\xc3' in file blend-If.py on line 6, but no encoding declared; see http://python.org/dev/peps/pep-0263/ for details

Add this as the second line:

Code: Select all

# -*- coding: UTF-8 -*-

This said, looking at the code, 1) it should be very slow (can be improved somewhat using "pixels regions" and 2) I wonder how different this is from this:

* Copy the two layers
* Blend them (top layer at 50% opacity then merge down)
* Apply over the original bottom layer in lighten only mode.

More accurate procedure:

* Duplicate both layers and apply "Desaturate > Luminosity" to each
* Set the top of these to "Grain extract" and merge
* Threshold the result to 128 or .5, this gives you a map of the pixels where the top layer is more luminous than the bottom one.
* In the Channels, do a "Channel to selection" on any of the RGB layers
* Back to the original top layer, add a layer mask and initialize to selection
* Set the opacity of that layer to 50%

Can of course be scripted, but on any large image it could be faster manually than running the script.

Great! It worked!
This UTF-8 line solved the problem!

The more accurate procedure is interesting. I will test it again in other editing situations.
But I can't compare it with the result of the code since it didn't work for me.

Anyway, you can see that the code didn't recreate the Blend-If as desired, since the preview makes a lot of difference to me.

teapot wrote:

Don't use spoiler tags, use code like this:

Code: Select all

blah
    indented blah

The code icon is two to the left of the spoiler icon.
Or you can just type the tags.

I saw this icon, but I thought it was for indentation or numeric list, so I didn't even click on it.

Code: Select all

Although I think I've used this code here before, I just don't remember if the icon image was the same.

teapot and ofnuts, thx a lot!

teapot · **Joined:** Dec 09, 2018 **Posts:** 736

ofnuts wrote:

More accurate procedure:

* Duplicate both layers and apply "Desaturate > Luminosity" to each
* Set the top of these to "Grain extract" and merge
* Threshold the result to 128 or .5, this gives you a map of the pixels where the top layer is more luminous than the bottom one.
* In the Channels, do a "Channel to selection" on any of the RGB layers
* Back to the original top layer, add a layer mask and initialize to selection
* Set the opacity of that layer to 50%

Hi ofnuts, Thank you for the procedure.

I was just wanting to try it but how do I apply "Desaturate > Luminosity"?

Under the menu Colours -> Desaturate -> Desaturate the modes I see are:
Luminance
Luma
Lightness (HSL)
Average (HSI Intensity)
Value (HSV)

ofnuts · **Joined:** Oct 25, 2010 **Posts:** 4812

teapot wrote:

ofnuts wrote:

More accurate procedure:

* Duplicate both layers and apply "Desaturate > Luminosity" to each
* Set the top of these to "Grain extract" and merge
* Threshold the result to 128 or .5, this gives you a map of the pixels where the top layer is more luminous than the bottom one.
* In the Channels, do a "Channel to selection" on any of the RGB layers
* Back to the original top layer, add a layer mask and initialize to selection
* Set the opacity of that layer to 50%

Hi ofnuts, Thank you for the procedure.

I was just wanting to try it but how do I apply "Desaturate > Luminosity"?

Under the menu Colours -> Desaturate -> Desaturate the modes I see are:
Luminance
Luma
Lightness (HSL)
Average (HSI Intensity)
Value (HSV)

Luminance is fine (actually the formula used in the script is called Luminance in some sources). In practice it shouldn't be too important since all these values increase in the same direction so comparing their luminance, lightness, or luma, shouldn't change the result.

teapot · **Joined:** Dec 09, 2018 **Posts:** 736

ofnuts wrote:

Luminance is fine (actually the formula used in the script is called Luminance in some sources). In practice it shouldn't be too important since all these values increase in the same direction so comparing their luminance, lightness, or luma, shouldn't change the result.

Thanks ofnuts :tyspin