For pango lineages we now sensibly order the lineages (rather than alphabetically) and colour them by partitioning into nextstrain clades and using a different colour scale for each clade.
This script depends on the python library pango_aliasor, however we are already using that in collapse-lineage-counts.py.
The viz library needs to be updated (e.g. https://github.com/nextstrain/forecasts-viz/pull/17) to parse the new colour definitions, which are stored in metdadata.variantColors, however older versions of the library will continue to work without error. The updated ordering should be immediately reflected by any version of the viz library.
I was unable to test this on my machine due to RuntimeError: jaxlib is version 0.4.6, but this version of jax requires version >= 0.4.11 which I think is because of issues with our arm64 docker image. I'll try to test via GitHub actions, which I'm presuming are x86...
Updates:
Test GitHub Action deployed to AWS; cloudwatch logs here. Job failed due to me not understanding how the model output filenames work. Will re-push in a minute.
For pango lineages we now sensibly order the lineages (rather than alphabetically) and colour them by partitioning into nextstrain clades and using a different colour scale for each clade.
This script depends on the python library
pango_aliasor
, however we are already using that incollapse-lineage-counts.py
.The viz library needs to be updated (e.g. https://github.com/nextstrain/forecasts-viz/pull/17) to parse the new colour definitions, which are stored in
metdadata.variantColors
, however older versions of the library will continue to work without error. The updated ordering should be immediately reflected by any version of the viz library.I was unable to test this on my machine due to
RuntimeError: jaxlib is version 0.4.6, but this version of jax requires version >= 0.4.11
which I think is because of issues with our arm64 docker image. I'll try to test via GitHub actions, which I'm presuming are x86...Updates: Test GitHub Action deployed to AWS; cloudwatch logs here. Job failed due to me not understanding how the model output filenames work. Will re-push in a minute.