Upgrade loquat dependency to refactor pipelines code

[laughedelic]

With https://github.com/ohnosequences/loquat/pull/58 it should be possible to define pipelines in a more convenient way (see that issue description).

[laughedelic]

@eparejatobes review this please. The main point of this refactoring is to reduce the boilerplate code in pipeline definitions. Check the test pipelines code for examples. Now all user has to define is

• MG7Parameters (in most scenarios using defaults for Illumina and PacBio)
• MG7Pipeline (optionally with Flash) which consists of
  • input data locations
  • location for outputs
  • common loquat settings (metadata, IAM role name, logs bucket)

Data mappings and loquat definitions are derived from this.

It is also possible to override config for each loquat and I'm in a doubt about it: probably it's better to define them always explicitly, because it's hard to set some sensible defaults for the workers group sizes. I still set them though (to the number of input samples). What do you think about it?

[eparejatobes]

Overall a nice improvement :+1:

I will add comments.

[eparejatobes]

In general in https://github.com/ohnosequences/mg7/blob/09170afb9cceddb9f8e8f00b50b888890d5ab226/src/main/scala/mg7/pipeline.scala the different components which make a config should be separated into different types.

[eparejatobes]

Done! also nice that now the flash part is separate :+1:

[laughedelic]

I think most comments are addressed. I'm moving to db.rna16s.

[eparejatobes]

OK fine. LGTM, the loquat part can wait.

[laughedelic]

Published as 1.0.0-M5-pr78-139-g91401e1