Remove hard dependency with S3

[pditommaso]

Summary

Currently Wave has an hard dependency with AWS S3 to store build logs and Conda lock files.

This prevents the use with other object storage system such as Google Object storage or Azure Blob containers.

The goal of this task is to remove such hard dependency, and allow storing logs and lock assets in other object storage system.

Goal

• Allow the use of Google Object storage to store logs and lock files
• Allow the use of Azure blob container to store logs and lock files

Non-goal

• Change the requirement for S3/R2 for the container blob caching component/service.

[pditommaso]

Notably the build logs service already rely on Micronaut abstraction to store those assets.

To implement this task is enough implementing a more flexible factory able to inspect the bucket schema e.g. s3, az or gs and create the corresponding client

wave/src/main/groovy/io/seqera/wave/service/aws/ObjectStorageOperationsFactory.groovy

Lines 47 to 53 in 456aa09

	@Singleton
	@Named("build-logs")
	ObjectStorageOperations<?, ?, ?> awsStorageOperations(@Named("DefaultS3Client") S3Client s3Client, InputStreamMapper inputStreamMapper) {
	AwsS3Configuration configuration = new AwsS3Configuration('build-logs')
	configuration.setBucket(storageBucket)
	return new AwsS3Operations(configuration, s3Client, inputStreamMapper)
	}

Bonus: This would also allow adding the support for local storage (ie. file:) for dev purposes

[pditommaso]

Solved by #808