ingest: deduplicate sequences using strain names

[joverlee521]

Context

Once we've completed #32, we can use strain names to deduplicate sequences.
This is necessary in case different groups sequence the same virus or if sequences are generated from different protocols.
(NOTE: This is separate from the versioning in GenBank, we already pull in the latest version of GenBank sequences).

Description

The duplicate sequences should probably be filtered out in a new script (e.g. ingest/bin/deduplicate-records) OR potentially use the augur deduplicate command (see nextstrain/augur#919).

We probably want to keep a file with all sequences in case people want the duplicate sequences for any reason.
The deduplicated files will be the main ones used for LAPIS and/or our monkeypox builds.

[jameshadfield]

Update: We currently have a duplicate in the hMPX build (MPXV-M5312_HM12_Rivers from accessions MT903340 and NC_063383). It’s not a huge problem as it's not in the current outbreak.