Reference selection improvements and nextclade generalization

[dpark01]

• viral-assemble 2.3.1.3 to 2.3.1.4 -- see release notes for viral-assemble but this changes default parameters to cluster genomes more aggressively and also rank orders hits by ANI * reference bases covered instead of just ANI so as to not favor tiny matches over long ones
• expose some skani parameters at the WDL task level
• update nextclade_one_sample to emit more useful (non-empty) output strings for multi-segment viruses like influenza
• add a taxid_to_nextclade workflow that takes a numeric NCBI taxid and converts it to a nextclade dataset name that can be used as the input value to nextclade_one_sample.dataset_name

[tomkinsc]

In a future PR we should say more about these single-character params, even if it's only to have entries in the parameter_meta section point to the skani docs.