Generalized advanced queries

[chaoran-chen]

For SARS-CoV-2, we have advanced variant queries (see #130 and docs). We should implement a generlized version.

The query format may but does not need to be exactly the same as for SARS-CoV-2. It is only important that, if they are not consistent, the existing SARS-CoV-2 implementation continues to exist and does not cause breaking changes to CoV-Spectrum.

The query format (of CoV-Spectrum) has three main components

• The logic operators: for SC2, it's: &, |, !, n-of and exactly-n-of
• The mutation selectors: for SC2, it's just the mutations in our usual syntax.
  • A mutation can be wrapped in a maybe().
• The metadata selectors:
  • For SC2, we have a quite special implementation, where Pango lineages are just written as is (e.g., B.1.1.7). Pango lineages called by Nextclade are prefixed (e.g., nextcladePangoLineage:B.1.1.7) as are Nextstrain clades (e.g., nextstrainClade:22B). There is no support for other metadata columns. Sub-lineages of Pango lineages can be queried (e.g., nextcladePangoLineage:B.1.1.7*)
  • For the new version, it would be nice to be able to support any metadata column.

Subtasks:

• Generalized advanced queries: Make suggestions for the syntax #1045
• Generalized advanced queries: Analyze C++ implementation #1046