Upgrade pangolin to v4

[jgadling]

Description of proposed changes

Upgrade pangolin (with UShER support) to version 4.1.3

I've done this by creating another intermediate-build image for UShER. The UShER team is publishing an image to dockerhub here, but it's based on an OS with different shared-library versions than the one we're using, so it's difficult/impractical to just use the built binaries from the upstream image.

I can easily refactor this to skip adding a third upstream image repository if that's a problem.

Related issue(s)

Testing

I spot-checked that pangolin works in UShER and scorpio mode in the final image.

UShER mode:

nextstrain:/tmp $ pangolin test_genome.fasta
****
Pangolin running in usher mode.
****
<content omitted>

Using UShER as inference engine.
****
Output file written to: /tmp/lineage_report.csv

PangoLEARN mode

nextstrain:/tmp $ pangolin --analysis-mode fast test_genome.fasta
****
Pangolin running in pangolearn mode.
****
Warning: pangoLEARN mode may use a significant amount of RAM, be aware that it will not suit every system.
Maximum ambiguity allowed is 0.3.
****
Query file:     /tmp/test_genome.fasta
****
<content omitted>
Processing block of 3 sequences 12/12/2022, 18:03:03
Complete 12/12/2022, 18:03:03
****
Output file written to: /tmp/lineage_report.csv

• Checks pass

[tsibley]

@jgadling Thanks for this proposal. I've held off on reviewing/testing this because I think we'd rather drop Pangolin/PangoLEARN all together if we can—rather than upgrade it—per discussion started in nextstrain/conda-base#7. We haven't resolved that discussion yet, but we should try to do so one way or another soon.