Standardize auxiliary input files in a "defaults" directory

This commit renames the `config` directory to `defaults` in both the ingest
phylogenetic workflow. The `defaults.yaml` for configuraiton values is also
renamed to `config.yaml` to be consistent with the pathogen-repo-guide.

https://github.com/nextstrain/pathogen-repo-guide/tree/main/ingest/defaults

ingest/Snakefile

@@ -5,7 +5,7 @@ min_version(
 )  # Snakemake 7.7.0 introduced `retries` directive used in fetch-sequences
 
 # Use default configuration values. Override with Snakemake's --configfile/--config options.
-configfile: "config/defaults.yaml"
+configfile: "defaults/config.yaml"
 
 
 send_slack_notifications = config.get("send_slack_notifications", False)

ingest/config/defaults.yaml → ingest/defaults/config.yaml

@@ -40,7 +40,7 @@ curate:
   geolocation_rules_url: 'https://raw.githubusercontent.com/nextstrain/ncov-ingest/master/source-data/gisaid_geoLocationRules.tsv'
   # The path to the local geolocation rules within the pathogen repo
   # The path should be relative to the ingest directory.
-  local_geolocation_rules: 'config/geolocation-rules.tsv'
+  local_geolocation_rules: 'defaults/geolocation-rules.tsv'
   # List of field names to change where the key is the original field name and the value is the new field name
   # The original field names should match the ncbi_datasets_fields provided above.
   # This is the first step in the pipeline, so any references to field names in the configs below should use the new field names
@@ -85,7 +85,7 @@ curate:
   authors_default_value: '?'
   # Path to the manual annotations file
   # The path should be relative to the ingest directory
-  annotations: "config/annotations.tsv"
+  annotations: "defaults/annotations.tsv"
   # The ID field in the metadata to use to merge the manual annotations
   annotations_id: 'genbank_accession'
   # The ID field in the metadata to use as the sequence id in the output FASTA file

phylogenetic/Snakefile

@@ -1,4 +1,4 @@
-configfile: "config/config_zika.yaml"
+configfile: "defaults/config_zika.yaml"
 
 rule all:
     input:

phylogenetic/rules/annotate_phylogeny.smk

@@ -54,7 +54,7 @@ rule translate:
     input:
         tree = "results/tree.nwk",
         node_data = "results/nt_muts.json",
-        reference = "config/zika_reference.gb"
+        reference = "defaults/zika_reference.gb"
     output:
         node_data = "results/aa_muts.json"
     shell:

phylogenetic/rules/export.smk

@@ -34,9 +34,9 @@ rule export:
         traits = "results/traits.json",
         nt_muts = "results/nt_muts.json",
         aa_muts = "results/aa_muts.json",
-        colors = "config/colors.tsv",
-        auspice_config = "config/auspice_config.json",
-        description = "config/description.md"
+        colors = "defaults/colors.tsv",
+        auspice_config = "defaults/auspice_config.json",
+        description = "defaults/description.md"
     output:
         auspice_json = "results/raw_zika.json",
         root_sequence = "results/raw_zika_root-sequence.json",

phylogenetic/rules/prepare_sequences.smk

@@ -60,7 +60,7 @@ rule filter:
     input:
         sequences = "data/sequences_all.fasta",
         metadata = "data/metadata_all.tsv",
-        exclude = "config/dropped_strains.txt",
+        exclude = "defaults/dropped_strains.txt",
     output:
         sequences = "results/filtered.fasta"
     params:
@@ -90,7 +90,7 @@ rule align:
     """
     input:
         sequences = "results/filtered.fasta",
-        reference = "config/zika_reference.gb"
+        reference = "defaults/zika_reference.gb"
     output:
         alignment = "results/aligned.fasta"
     shell: