Fix malt tests that are using a 739.0 MiB zip as input

[maxulysse]

          so the input for unzip on line 19 is a `739.0 MiB` zip...

I'm guessing that's why the tests are failing...

Originally posted by @maxulysse in #5897 (comment)

[jfy133]

Currently this is required for the module to test.

The actual error reported in that CI was:

    > ERROR ~ Access Denied (Service: Amazon S3; Status Code: 403; Error Code: AccessDenied; Request ID: XD54JYBDPDWDSX0E; S3 Extended Request ID: JmDS4i8VpQQTDw/5SrGhY7knZ3AJxsxKCfWID2bGjrMZrmvZn/3cZ/FjOIKuvwnBCJPOGxTfEao=; Proxy: null)

I'm wondering now if I copied the required taxdmp files to our bucket somewhere...

[jfy133]

Ok indeed I did:

   input[0] = [[], file("s3://ngi-igenomes/test-data/createtaxdb/taxonomy/megan-nucl-Feb2022.db.zip", checkIfExists: true)]

@nf-core/infrastructure can you think why reason why GHA can't access the igenomes bucket?

[bebosudo]

That file seems to be openly available, both with aws cli

[alberto@fedora]$ podman run --rm -it docker.io/amazon/aws-cli s3 --no-sign-request cp s3://ngi-igenomes/test-data/createtaxdb/taxonomy/megan-nucl-Feb2022.db.zip /tmp

and as http https://ngi-igenomes.s3.amazonaws.com/test-data/createtaxdb/taxonomy/megan-nucl-Feb2022.db.zip

Note the --no-sign-request, otherwise you get a unable to locate credentials error, see this thread: aws/aws-cli#904

I don't know how you're retrieving that file though, if with something different than the aws cli, you may want to check the current role/user you're assuming, as that may have some restrictions that prevent it from reaching that bucket and object.

[jfy133]

Hrm, must be something nextflow side then...

[ewels]

This sounds like the same "just rm -r ~/.aws" answer that we give on slack on a regular basis.

Are the CI tests configuring AWS access for anything else?

[maxulysse]

Not that I'm aware

[jfy133]

This sounds like the same "just rm -r ~/.aws" answer that we give on slack on a regular basis.

Are the CI tests configuring AWS access for anything else?

Ooff, ok. So we should add that as an upstream step of the actual test in the gha?