Merge pull request #1161: Use weighted sampling for recent samples in…

… global builds
nextstrain · Dec 17, 2024 · 2f32f78 · 2f32f78
2 parents 9adce5f + 941613b
commit 2f32f78
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 0 deletions.
diff --git a/nextstrain_profiles/nextstrain-gisaid-21L/builds.yaml b/nextstrain_profiles/nextstrain-gisaid-21L/builds.yaml
@@ -294,6 +294,7 @@ subsampling:
       max_date: "--max-date 1M"
     recent:
       group_by: "country week"
+      group_by_weights: "defaults/population_weights.tsv"
       max_sequences: 4100
       min_date: "--min-date 1M"
 
@@ -308,6 +309,7 @@ subsampling:
       max_date: "--max-date 2M"
     recent:
       group_by: "country week"
+      group_by_weights: "defaults/population_weights.tsv"
       max_sequences: 4100
       min_date: "--min-date 2M"
 
@@ -322,6 +324,7 @@ subsampling:
       max_date: "--max-date 6M"
     recent:
       group_by: "country year month"
+      group_by_weights: "defaults/population_weights.tsv"
       max_sequences: 4100
       min_date: "--min-date 6M"
 

diff --git a/nextstrain_profiles/nextstrain-gisaid/builds.yaml b/nextstrain_profiles/nextstrain-gisaid/builds.yaml
@@ -286,6 +286,7 @@ subsampling:
       max_date: "--max-date 1M"
     recent:
       group_by: "country week"
+      group_by_weights: "defaults/population_weights.tsv"
       max_sequences: 4100
       min_date: "--min-date 1M"
 
@@ -300,6 +301,7 @@ subsampling:
       max_date: "--max-date 2M"
     recent:
       group_by: "country week"
+      group_by_weights: "defaults/population_weights.tsv"
       max_sequences: 4100
       min_date: "--min-date 2M"
 
@@ -314,6 +316,7 @@ subsampling:
       max_date: "--max-date 6M"
     recent:
       group_by: "country year month"
+      group_by_weights: "defaults/population_weights.tsv"
       max_sequences: 4100
       min_date: "--min-date 6M"
 

diff --git a/nextstrain_profiles/nextstrain-open/builds.yaml b/nextstrain_profiles/nextstrain-open/builds.yaml
@@ -286,6 +286,7 @@ subsampling:
       max_date: "--max-date 1M"
     recent:
       group_by: "country week"
+      group_by_weights: "defaults/population_weights.tsv"
       max_sequences: 4100
       min_date: "--min-date 1M"
 
@@ -300,6 +301,7 @@ subsampling:
       max_date: "--max-date 2M"
     recent:
       group_by: "country week"
+      group_by_weights: "defaults/population_weights.tsv"
       max_sequences: 4100
       min_date: "--min-date 2M"
 
@@ -314,6 +316,7 @@ subsampling:
       max_date: "--max-date 6M"
     recent:
       group_by: "country year month"
+      group_by_weights: "defaults/population_weights.tsv"
       max_sequences: 4100
       min_date: "--min-date 6M"