From 3d8bed9885feae7d55851c72fd15269b5af95567 Mon Sep 17 00:00:00 2001
From: Trevor Bedford <trevor@bedford.io>
Date: Fri, 26 Jul 2024 13:15:23 -0700
Subject: [PATCH] Allow build-specific --narrow-bandwidth param in frequencies

There was a bug in the current behavior where the rule `frequencies` was calling `config["frequencies"]["narrow_bandwidth"]`. This resulted in --narrow-bandwidth to be fixed to whatever was specified in parameters.yaml. However, we were relying on build-specific settings to differentiate behavior in the all-time vs 6m vs 2m vs 1m builds, ala:

```
frequencies
    global_1m:
      narrow_bandwidth: 0.019
```

This commit fixes this issue and allows overriding of narrow_bandwidth in parameters.yaml by build-specific settings. It also provides a genuine default (equal to augur default) so that parameters.yaml doesn't have to always specify narrow_bandwidth.
---
 .../nextstrain-gisaid/builds.yaml             | 29 -------------------
 workflow/snakemake_rules/common.smk           | 11 +++++++
 workflow/snakemake_rules/main_workflow.smk    |  2 +-
 3 files changed, 12 insertions(+), 30 deletions(-)

diff --git a/nextstrain_profiles/nextstrain-gisaid/builds.yaml b/nextstrain_profiles/nextstrain-gisaid/builds.yaml
index ab07dcc63..cdb06d629 100644
--- a/nextstrain_profiles/nextstrain-gisaid/builds.yaml
+++ b/nextstrain_profiles/nextstrain-gisaid/builds.yaml
@@ -890,116 +890,87 @@ frequencies:
   reference:
     min_date: "2020-01-01"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   global_1m:
     min_date: "1M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   global_2m:
     min_date: "2M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   global_6m:
     min_date: "6M"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   global_all-time:
     min_date: "2020-01-01"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   africa_1m:
     min_date: "1M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   africa_2m:
     min_date: "2M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   africa_6m:
     min_date: "6M"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   africa_all-time:
     min_date: "2020-01-01"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   asia_1m:
     min_date: "1M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   asia_2m:
     min_date: "2M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   asia_6m:
     min_date: "6M"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   asia_all-time:
     min_date: "2020-01-01"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   europe_1m:
     min_date: "1M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   europe_2m:
     min_date: "2M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   europe_6m:
     min_date: "6M"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   europe_all-time:
     min_date: "2020-01-01"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   north-america_1m:
     min_date: "1M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   north-america_2m:
     min_date: "2M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   north-america_6m:
     min_date: "6M"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   north-america_all-time:
     min_date: "2020-01-01"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   oceania_1m:
     min_date: "1M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   oceania_2m:
     min_date: "2M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   oceania_6m:
     min_date: "6M"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   oceania_all-time:
     min_date: "2020-01-01"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   south-america_1m:
     min_date: "1M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   south-america_2m:
     min_date: "2M"
     narrow_bandwidth: 0.019
-    recent_days_to_censor: 7
   south-america_6m:
     min_date: "6M"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
   south-america_all-time:
     min_date: "2020-01-01"
     narrow_bandwidth: 0.038
-    recent_days_to_censor: 7
diff --git a/workflow/snakemake_rules/common.smk b/workflow/snakemake_rules/common.smk
index 0c2713950..022bdd8a7 100644
--- a/workflow/snakemake_rules/common.smk
+++ b/workflow/snakemake_rules/common.smk
@@ -207,6 +207,17 @@ def _get_max_date_for_frequencies(wildcards):
             datetime.date.today() - offset
         )
 
+def _get_narrow_bandwidth_for_wildcards(wildcards):
+    # check if builds.yaml contains frequencies:{build_name}:narrow_bandwidth
+    if wildcards.build_name in config["frequencies"] and 'narrow_bandwidth' in config["frequencies"][wildcards.build_name]:
+        return config["frequencies"][wildcards.build_name]["narrow_bandwidth"]
+    # check if parameters.yaml contains frequencies:narrow_bandwidth
+    elif "frequencies" in config and "narrow_bandwidth" in config["frequencies"]:
+        return config["frequencies"]["narrow_bandwidth"]
+    # else return augur frequencies default value
+    else:
+        return 0.0833
+
 def _get_upload_inputs(wildcards):
     # Do whatever the configuration says if it has opinions.
     if "upload" in config:
diff --git a/workflow/snakemake_rules/main_workflow.smk b/workflow/snakemake_rules/main_workflow.smk
index 2bd018d25..10c19542f 100644
--- a/workflow/snakemake_rules/main_workflow.smk
+++ b/workflow/snakemake_rules/main_workflow.smk
@@ -1164,7 +1164,7 @@ rule tip_frequencies:
         max_date = _get_max_date_for_frequencies,
         pivot_interval = config["frequencies"]["pivot_interval"],
         pivot_interval_units = config["frequencies"]["pivot_interval_units"],
-        narrow_bandwidth = config["frequencies"]["narrow_bandwidth"],
+        narrow_bandwidth = _get_narrow_bandwidth_for_wildcards,
         proportion_wide = config["frequencies"]["proportion_wide"]
     resources:
         # Memory use scales primarily with the size of the metadata file.