Add Newick tree support

[chrisvanrun]

Part of the pitch:

• https://github.com/DIAGNijmegen/rse-roadmap/issues/354

This PR adds support for Newick files.

• Validates content by attempting to parse the trees.
• It forces .newick suffix in the relative path.
• Uploader (Uppy) allows selection of:
  • *.nwk, used by the test files in the parser library.
  • *.tree, defined by Wikipedia.
  • *.newick, provided in the example files.

At this point, it runs on acks-late-micro-short. As the Newick are not very hard to parse or very big.

[jmsmkn]

acks-late-micro-short really doesn't have much memory available, so don't be surprised when we get memory errors from that validation.

[chrisvanrun]

acks-late-micro-short really doesn't have much memory available, so don't be surprised when we get memory errors from that validation.

Would you agree that for a first implementation, it is OK? They should be around the same size as the multiple annotation JSON types. I'll be looking at re-routing file-to-object task to larger workers when implementing BIOM.

[jmsmkn]

Yes totally fine, was just a heads up.

[chrisvanrun]

Summary of updates:

The mime-types were not being really used as MIME-types but rather as input to Uppy. At the same time the need to support .newick as file extension and uploads caused several other properties. This added complexity.

Since we don't need to actually map to mime types I've refactored things to a allowed_file_types map and lookup, combining both MIME and the other file type definitions Uppy supports: https://uppy.io/docs/uppy/#restrictions.

Newick validator now also handles time-out errors, memory errors. It also no longer exposes generic exceptions.