GenoFLU for fauna datasets

[jameshadfield]

Generates and exports both the per-segment GenoFLU results and the whole-genome genotype/constellation.

There's a few outstanding to-dos, to be done either here or elsewhere:

• If there are fewer than 8 segments then genoflu doesn't report this in the results (or the log). We already have n_segments in the metadata so we could use this to generate a "<8 segments sequenced" label if desired.
• What's more frustrating for samples with fewer than 8 segments is that we don't get the per-segment calls for the sequenced segments. Presumably we could modify genoflu to report these.
• For genomes where some of the segments are too diverged my current parsing approach doesn't pull them out. While it was fun to do everything in csvtk we should switch to a python script to correctly parse annotations where (e.g.) the genotype list is MP:ea3, HA:ea3, NA:ea5, PB1:ea3 and the genotype is Not assigned: Only 4 segments >98.0% match found of total 8 segments in input file

[jameshadfield]

I'm going to merge this now as it works locally and will enable us to improve the cattle-outbreak segment-level filtering discussed here. I'll run actions to re-ingest fauna and then update the datasets and update this comment accordingly.

• action to ingest fauna completed successfully and all 9 genoflu columns are now in the S3 metadata.
• action to re-run all (fauna-based) trees RUNNING