default nucleotide and amino acid sequence

[fengelniederhammer]

resolves #454

Summary

Changes in this PR:

• when no default nucleotide sequence is set in the config, set it to absent (instead of defaulting to "main")
• read the default amino acid sequence from the config (instead of hardcoding it to absent)
• validate that the values are actually present in the reference genomes (if set)
• if there is only one sequence of a type in the reference genomes, also set the default sequence in the database config
• add a unit test that checks that relevant mutation queries work without a sequenceName field, if the config has the default sequences.

PR Checklist

• [ ] All necessary documentation has been adapted or there is an issue to do so. -> Docs: default sequences in database config LAPIS#862
• The implemented feature is covered by an appropriate test.

[github-actions]

This is a preview of the changelog of the next release:

0.2.9 (2024-07-18)

Features

• have no default sequence by default, implement default amino acid sequence from config (448a6de), closes #454
• set the default sequence when there is only a single sequence in the reference genomes (092c65b), closes #454
• throw error in preprocessing when default sequence is not in reference genomes (c739ad6), closes #454

Bug Fixes

• streaming: report an error when the unimplemented offset or limit matters (3be9c06), closes #511

[Taepper]

That was quick! A few minor comments

[Taepper]

I find this function name a little confusing

[fengelniederhammer]

Because of the name or because of the content? I'm also not super happy with the name. Suggestions are welcome ;)

[Taepper]

The name made me think that it would not mutate the config
Maybe finalize / resolve?

Otherwise an extra class ValidatedDatabaseConfig which also takes the ReferenceGenomes in its constructor would be viable

or we can leave it as a TODO for #426 , where we want to specify the unaligned sequences (and I would also say the aligned nuc and aa sequences) in the database config to reduce this dependency. Then the reference genomes is no longer a config for the sequences in the database but only a map for looking up the references

[fengelniederhammer]

Ok good point. check doesn't imply mutating. I renamed it to finalizeConfig.

ValidatedDatabaseConfig is only partially true, because you also need to validate it against the data (e.g. that all metadata fields are present etc.). So I'm not sure whether that would be a good name either.
If we want to redesign this anyway, then I'd propose to leave the rest to #426.