NCBI Taxonomy download snakefile

config.yaml

@@ -8,12 +8,15 @@ ALVISNLP: '~/code/alvisnlp/.test/alvisnlp/bin/alvisnlp'
 REWRITE_TAXONOMY: '~/code/bibliome-java-utils/test/install/bin/rewrite-taxonomy'
 
 
-# NCBI Taxonomy files
-NCBI_DIR: 'ncbi-taxonomy_2021-03-26'
-
-
 # Output and working directories
 OUTDIR: 'test'
 
+
+
+
+
+
 DSMZ_STRAINS_DIR: 'dsmz-strains'
 DSMZ_MATCH_DIR: 'dsmz-match'
+NCBI_DIR: 'ncbi-taxonomy'
+NCBI_ZIP_URL: 'https://ftp.ncbi.nlm.nih.gov/pub/taxonomy/taxdmp.zip'

dsmz-match.snakefile

 configfile: 'config.yaml'
 
-import glob
 
 rule match:
     '''

ncbi-download.snakefile

+configfile: 'config.yaml'
+
+
+rule unzip:
+    output:
+        config['OUTDIR'] + '/' + config['NCBI_DIR'] + '/nodes.dmp'
+
+    input:
+        config['OUTDIR'] + '/' + config['NCBI_DIR'] + '/taxdmp.zip'
+
+    shell:
+        '''unzip -d {config[OUTDIR]}/{config[NCBI_DIR]} {input}'''
+
+rule download:
+    output:
+        config['OUTDIR'] + '/' + config['NCBI_DIR'] + '/taxdmp.zip'
+
+    shell:
+        '''curl -o {output} '{config[NCBI_ZIP_URL]}' '''