HI!
I am trying to inject a series of values sequentially into a tool so that it performs different cropping operations on an image. I have a tabular file that contains the following data:
| Column 1 | Column 2 | Column 3 | Column 4 |
|---|---|---|---|
| 130 | 124 | 100 | 200 |
| 710 | 400 | 50 | 60 |
| 330 | 730 | 110 | 120 |
Columns 1 and 2 represent the x and y coordinates on the image, and columns 3 and 4 represent the width and height of the crop. Each row defines a rectangular region that needs to be cropped.
I need to inject these four parameters to the “Convert image format with BioFormats tool” which can generate a cropped image. The desired output is three cropped images, one for each row of the table.
At the moment, I extract each column, split the resulting list by group, and then parse the parameters (which produces a list of three datasets). This list is then injected into the Convert image format tool. However, I have two issues:
-
The values are injected into the conversion tool in a random order. All four values from a single row should be supplied to the same job, but the “Split by group” step assigns dataset IDs arbitrarily, so the rows get mixed up.
-
I am not sure that “Split by group” is the right approach for this workflow. If two rows contain identical coordinates, they will be grouped together resulting in unpaired datasets.
I would appreciate any advice!
Attached are a sample history and the workflow: