GISAID build fails due to new "country" Lugansk missing from weights

[corneliusroemer]

In CI, test build fails:

[batch] [2025-12-11T12:26:59+00:00] ================================================================================
[batch] [2025-12-11T12:26:59+00:00] Sampling with weights defined by defaults/population_weights.tsv.
[batch] [2025-12-11T12:26:59+00:00] ERROR: The input metadata contains these values under the following columns that are not covered by 'defaults/population_weights.tsv':
[batch] [2025-12-11T12:26:59+00:00] - 'country': ['Lugansk']
[batch] [2025-12-11T12:26:59+00:00] To fix this, either:
[batch] [2025-12-11T12:26:59+00:00] (1) specify weights explicitly - add entries to 'defaults/population_weights.tsv' for the values above, or
[batch] [2025-12-11T12:26:59+00:00] (2) specify a default weight - add an entry to 'defaults/population_weights.tsv' with the value 'default' for all columns

https://github.com/nextstrain/ncov/actions/runs/20131845493/job/57775040965#step:5:8097

[victorlin]

Lugansk is a city not country, so this is an issue with the metadata from ncov-ingest.

[corneliusroemer]

It's an oblast of Ukraine that's occupied by Russia and GISAID bypasses such conflicts by creating new top level entities.

We might want to stick to internationally recognized countries and add rewrite rules to put Lugansk under Ukraine (as "Ukraine/Luhansk")

[Walde1nsamkeit]

On GISAID, there are both sequences from "Europe / Ukraine / Severodonetsk" uploaded by Ukraine and sequences from "Europe / Lugansk / Severodonetsk" uploaded by Russian occupation government.
Formerly some sequences from Russian occupied Crimea is uploaded with the location displayed as "Europe / Russia / Crimea" and being revised into "Europe / Crimea" by GISAID later.
Now if you type "Europe / Russia / Crimea" into the location textbox, it will be revised into "Europe / Crimea". (The correct form should be "Europe / Ukraine / Crimea" or "Europe / Ukraine / Krym".)