Pango build for Nextclade

[corneliusroemer]

Plan:

• [x] Join designations onto metadata
• [x] Find good subsampling strategy: focus on recent diversity, big parent lineages with lots of diversity like B.1.617.2 should get more sequences than narrow ones like AY.39.1.1

Not blocking but important for good coverage with lineages from countries with whitespace in their names:

• [ ] Join using canonicalized keys
• [ ] For Nextclade no timetree -> how does it scale with size?

resolves #14

[corneliusroemer]

Test build: https://nextstrain.org/groups/neherlab/ncov/pango

[corneliusroemer]

Bug in preprocess on this branch, probably due to lack of key in (default) config

[corneliusroemer]

Mapping designation strain names to Nextstrain strain names

• Select only columns strain name and GISAID EPI ISL from metadata
• Normalize meta strain name down to alphanumeric -> mapping file
• Normalize pango strain name to alphanumeric, then use mapping to translate to EPI ISL=
• Create new pango designations file with GISAID EPI ISL for joining later