Added op3loader

[Olga013]

The continuation of Paulos' PR:
#921

[szalata]

Thanks for the great work!
I have added comments about a few things that need to be fixed, primarily op3_loader.sh

[szalata]

this had to be fixed/removed

[szalata]

this is about another dataset

[szalata]

this concerns another dataset, needs to be updated with information from our dataset, e.g. from the perturbation prediction repo

[szalata]

lgtm! Thank you for adding it!
@rcannood unless you see more issues, we could test it and merge

[Olga013]

@rcannood, I also have some questions. Would it be possible for you to clarify them?

1. I had a problem with normalization_methods, I did not manage to choose one when I specify them in the command line. Every time I got a default set of them containing 4 or 5 methods:
   https://github.com/Olga013/openproblems/blob/e9618766a374f1e744fd9292d4ef4a6f96f46d96/src/datasets/resource_scripts/op3_loader_test.sh#L23

   However, I managed to choose one method by creating and using a params file like this: https://github.com/Olga013/openproblems/blob/e9618766a374f1e744fd9292d4ef4a6f96f46d96/src/datasets/resource_scripts/op3_loader.sh#L16

   Do you have some suggestions, why this may happen?

2. I found that the Nextflow script related to process_op3 workflow contained the function extract_metadata. But when I ran this workflow it did not work until I changed the function name from extract_metadata to extract_uns_metadata. In addition, I found that other alternative scripts contain the extract_metadata function (e.g. extract_metadata in the process_openproblems_v1 workflow).

   Do we need to change the rest of them to extract_uns_metadata?

3. Do we need to add the sequence of commands for running the loader and processor workflows to the README.md file?

4. Which the Nextflow config file is better to use (in which cases we refer to nextflow_helpers and in which cases we create our own file)? common/nextflow_helpers/labels_ci.config or /tmp/nextflow.config.

5. Do we need to specify the requirements on the site that the version of the Nextflow should not be higher than 25.02.0-edge because of this issue? (I used 25.01.0-edge)

[rcannood]

I had a problem with normalization_methods, I did not manage to choose one when I specify them in the command line. Every time I got a default set of them containing 4 or 5 methods:
https://github.com/Olga013/openproblems/blob/e9618766a374f1e744fd9292d4ef4a6f96f46d96/src/datasets/resource_scripts/op3_loader_test.sh#L23

Not sure what went wrong here, but bash commands shouldn't include colons.

So this:

  --dataset_reference: GSE279945 \
  --normalization_methods: log_cp10k \

Should be:

  --dataset_reference GSE279945 \
  --normalization_methods log_cp10k \

2. I found that the Nextflow script related to process_op3 workflow contained the function extract_metadata. But when I ran this workflow it did not work until I changed the function name from extract_metadata to extract_uns_metadata. In addition, I found that other alternative scripts contain the extract_metadata function (e.g. extract_metadata in the process_openproblems_v1 workflow).

Do we need to change the rest of them to extract_uns_metadata?

Indeed, this should be extract_uns_metadata!

3. Do we need to add the sequence of commands for running the loader and processor workflows to the README.md file?

Sure, that would help!

4. Which the Nextflow config file is better to use (in which cases we refer to nextflow_helpers and in which cases we create our own file)? common/nextflow_helpers/labels_ci.config or /tmp/nextflow.config.

Depends on the needs of the dataset loader. If labels_ci.config works, then use that :)

5. Do we need to specify the requirements on the site that the version of the Nextflow should not be higher than 25.02.0-edge because of this issue? (I used 25.01.0-edge)

This can be solved by updating viash_version in _viash.yaml to 0.9.4 !

[Olga013]

@rcannood, @szalata I think that PR is ready to merge.