Add background_removal tool (skimage)

[rmassei]

Hi!
I think it would be nice to have in the collection a tool to automatically perform background removal from an intensity image.
This tool is based on skimage and more info can be found here.

At the present, three potential options:

1. Rolling-Ball Algorithm
2. Difference of Gaussians
3. Top-Hat Filter

Original

Rolling-Ball

Difference of Gaussians

Output is a 3 channel, 16 bit TIFF

---

FOR THE CONTRIBUTOR — Please fill out if applicable

Please make sure you have read the CONTRIBUTING.md document (last updated: 2024/04/23).

• License permits unrestricted use (educational + commercial).

If this PR adds or updates a tool or tool collection:

• This PR adds a new tool or tool collection.
• This PR updates an existing tool or tool collection.
• Tools added/updated by this PR comply with the Naming and Annotation Conventions for Tools in the Image Community in Galaxy (or explain why they do not).

[kostrykin]

Thanks a lot for this work!

There are a few comments below.

Also, please feel free to add yourself as a creator, as in:

galaxy-image-analysis/tools/bfconvert/bfconvert.xml

Line 10 in 5a1ff1a

<expand macro="creators/alliecreason"/>

[bgruening]

Input staging problem: Test input file (img.png) cannot be found.

You can find the planemo output on this site for example: https://github.com/BMCV/galaxy-image-analysis/actions/runs/9907411219

[rmassei]

Accepted all the changes and added the macro tests

[kostrykin]

Thanks @rmassei!

There are a few more comments below.

[kostrykin]

So I think there was a misunderstanding about my previous comment :)
#125 (comment)

I mean, the issue that the range of values is changed by the tool remains, even after you changed the test input data to an uint16 image. We need to dig a bit deeper here.

First, lets see how the wrapped functions change the data type:

• skimage.filters.difference_of_gaussians changes everything to float.
• skimage.morphology.white_tophat does not change the data type.
• skimage.restoration.rolling_ball does not change the data type either.

So we only need to care for conversion to float. I suggest something like the above, but maybe it can be done more elegantly.

[rmassei]

thanks @kostrykin, I did not get the problem in the first place.
Test were updated and it seems that the new function is working by correctly converting to the input data type.

[kostrykin]

Thanks @rmassei, I think we are almost done :) only a few minor comments below.

And there are a few flake8 linting issues which need to be addressed:
https://github.com/BMCV/galaxy-image-analysis/actions/runs/9937226470/job/27448034390?pr=125#step:6:1

tools/background_removal/background_removal.py:28:1: W293 blank line contains whitespace
tools/background_removal/background_removal.py:29:1: E302 expected 2 blank lines, found 1
tools/background_removal/background_removal.py:29:53: W291 trailing whitespace
tools/background_removal/background_removal.py:30:8: W291 trailing whitespace
tools/background_removal/background_removal.py:31:63: W291 trailing whitespace
tools/background_removal/background_removal.py:32:8: W291 trailing whitespace
tools/background_removal/background_removal.py:33:42: W291 trailing whitespace
tools/background_removal/background_removal.py:34:21: W291 trailing whitespace
tools/background_removal/background_removal.py:35:41: W291 trailing whitespace
tools/background_removal/background_removal.py:36:48: W291 trailing whitespace
tools/background_removal/background_removal.py:37:42: W291 trailing whitespace
tools/background_removal/background_removal.py:38:47: W291 trailing whitespace
tools/background_removal/background_removal.py:39:41: W291 trailing whitespace
tools/background_removal/background_removal.py:40:46: W291 trailing whitespace
tools/background_removal/background_removal.py:41:10: W291 trailing whitespace
tools/background_removal/background_removal.py:42:79: W291 trailing whitespace

[kostrykin]

Much thanks!

[kostrykin]

https://github.com/BMCV/galaxy-image-analysis/actions/runs/9937821722?pr=125

pillow and tifffile are not installed, but required to compare files using the 'image_diff' method

I think this is due to #124
where we removed python -m pip install pillow tifffile in 6bc8575

@bernt-matthias Any idea how to fix this? Shall we add it back again?

[rmassei]

yes, I got the same error locally before running python -m pip install pillow tifffile

[kostrykin]

Could not update background_removal
Failed to find repository for owner/name ufz/background_removal
Repository background_removal updated successfully.
Could not update background_removal
Failed to find repository for owner/name ufz/background_removal
Unexpected HTTP status code: 400: {"err_msg": "You already own a repository named background_removal, please choose a different name.", "err_code": 400008}
Repository [background_removal] does not exist in the targeted Tool Shed.
Failed to update metadata for repository 'background_removal' on the main Tool Shed.
Repository contents updated but failed to update metadata for repository 'background_removal' on the main Tool Shed.

https://github.com/BMCV/galaxy-image-analysis/actions/runs/9946388590/job/27477147950#step:6:222

Not sure what this error means. @bgruening do you know?

On the one hand it says that background_removal repository is already owned, on the other hand it complains that it doesn't exist… shouldn't it create it?

[bgruening]

the owner is wrong, we can only push to imgteam

[bgruening]

Please change the owner in the shed.yml file