JensWendt
commented
2 years ago Hi @will-moore ,
I think we should orient ourselves at the same standard that the KV_from_csv script implements.
For every object with KV-pairs one row.
For every Key one column.
Resulting in this output for your above sample:
| well | image | key1 | key2 |
|---|---|---|---|
| A1 | value1 | value2 | |
| A1 | field1 | value10 | value20 |
| A1 | field2 | value11 | value22 |
In general I assume it is highly unlikely that a well and the "underlying" images have differing KV-pairs. I personally recommend my users to use the same KV-pairs for both, for a better overview in certain situations. With the same reasoning it is unlikely that the image names are not unique in a plate and therefore the well name before the image might not be required in a single plate setup. But if we go for multiple plates in one .csv I think the whole "chain" plate-name, well-name and image-name is required.
Nevertheless I think that creating a .csv that would allow the recreation of the same annotation-setup if it immediately gets used for the reverse KV_from_csv is the goal.
I hope that helps (although it comes pretty late, sorry)
snoopycrimecop
I was prompted by the need for test data for https://github.com/ome/omero-scripts/pull/195, so I have updated KeyVal_to_csv.py to support
Plateinput. I seem to remember @abhamacher requested this too somewhere?I wasn't sure whether to export KV pairs from
wellsorimagesfor a Plate. Combining different KV pairs from images and wells could be tricky. My initial update allows the user to choose, with a boolean input "Export Wells"? for a Plate.But if we can define how to combine KV pairs from Wells and Images, I could give it a go.
E.g. if you have a Well containing 2 images:
A1with KV pairs:key1: value1, key2: value2field1with KV pairs:key1: value10, key2: value20field2with KV pairs:key1: value11, key2: value22What should the CSV look like? 1 row or 2 rows or even 3? cc @JensWendt