fix: pangolin data updates (#662)

* restructure data download * fmt * change name * change name of rule * remove old rules * change input name
IKIM-Essen · Aug 20, 2024 · f46ec83 · f46ec83
1 parent 6d78c8a
commit f46ec83
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 27 deletions.
diff --git a/workflow/rules/ref.smk b/workflow/rules/ref.smk
@@ -191,32 +191,15 @@ rule get_human_genome:
         "curl -SL -o {output} {params.human_genome} 2> {log}"
 
 
-rule update_pangoLEARN:
+rule get_pangolin_data:
     output:
-        directory("results/{date}/pangolin/pangoLEARN"),
+        "results/{date}/pangolin/pangolin-data.log",
     log:
-        "logs/{date}/pangolin/update.log",
+        "logs/{date}/pangolin/pangolin-data.log",
     conda:
-        "../envs/unix.yaml"
-    shell:
-        "(mkdir -p {output} &&"
-        " curl -L https://github.com/cov-lineages/pangoLEARN/archive/master.tar.gz |"
-        " tar xvz --strip-components=1 -C {output})"
-        " > {log} 2>&1"
-
-
-rule update_lineages:
-    output:
-        directory("results/{date}/pangolin/lineages"),
-    log:
-        "logs/{date}/pangolin/update.log",
-    conda:
-        "../envs/unix.yaml"
+        "../envs/pangolin.yaml"
     shell:
-        "(mkdir -p {output} &&"
-        " curl -L https://github.com/cov-lineages/lineages/archive/master.tar.gz | "
-        " tar xvz --strip-components=1 -C {output})"
-        " > {log} 2>&1"
+        "pangolin --update-data > {log} 2>&1 && cp {log} {output}"
 
 
 rule get_gisaid_provision:

diff --git a/workflow/rules/strain_calling.smk b/workflow/rules/strain_calling.smk
@@ -141,19 +141,16 @@ rule kallisto_plot_all_strains:
 rule pangolin_call_strains:
     input:
         contigs=get_pangolin_input,
-        pangoLEARN="results/{date}/pangolin/pangoLEARN",
-        lineages="results/{date}/pangolin/lineages",
+        data="results/{date}/pangolin/pangolin-data.log",
     output:
         "results/{date}/tables/strain-calls/{sample}.{stage}.strains.pangolin.csv",
     log:
         "logs/{date}/pangolin/{sample}.{stage}.log",
-    params:
-        pango_data_path=lambda w, input: os.path.dirname(input.pangoLEARN),
     conda:
         "../envs/pangolin.yaml"
     threads: 8
     shell:
-        "pangolin {input.contigs} --data {params.pango_data_path} --outfile {output} > {log} 2>&1"
+        "pangolin {input.contigs} --outfile {output} > {log} 2>&1"
 
 
 rule pangolin_plot_all_strains: