Merge pull request #251 from Ferlab-Ste-Justine/fix/clin-2536

fix: CLIN-2536 fix configs for new datasets
Ferlab-Ste-Justine · Dec 12, 2024 · 1e45acb · 1e45acb
2 parents d038195 + 97f7620
commit 1e45acb
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 13 deletions.
diff --git a/datalake-spark3/src/main/resources/reference_kf.conf b/datalake-spark3/src/main/resources/reference_kf.conf
@@ -43,16 +43,16 @@ datalake {
         },
         {
             format=VCF
-            id="raw_gnomad_genomes_v4"
+            id="raw_gnomad_genomes_v3"
             keys=[]
             loadtype=OverWrite
             partitionby=[]
-            path="/release/4.1/vcf/genomes/gnomad.genomes.v4.1.sites.chr[^M]*.vcf.bgz"
+            path="/release/3.1/vcf/genomes/gnomad.genomes.v3.1.sites.chr[^M]*.vcf.bgz"
             readoptions {
                 flattenInfoFields="true"
                 "split_multiallelics"="true"
             }
-            storageid=gnomadv4
+            storageid=gnomad
             writeoptions {
                 "created_on_column"="created_on"
                 "is_current_column"="is_current"
@@ -63,11 +63,11 @@ datalake {
         },
         {
             format=VCF
-            id="raw_gnomad_genomes_v3"
+            id="raw_gnomad_genomes_v4"
             keys=[]
             loadtype=OverWrite
             partitionby=[]
-            path="/release/3.1/vcf/genomes/gnomad.genomes.v3.1.sites.chr[^M]*.vcf.bgz"
+            path="/release/4.1/vcf/genomes/gnomad.genomes.v4.1.sites.chr[^M]*.vcf.bgz"
             readoptions {
                 flattenInfoFields="true"
                 "split_multiallelics"="true"
@@ -880,7 +880,7 @@ datalake {
             partitionby=[
                 chromosome
             ]
-            path="/public/gnomad_genomes_4"
+            path="/public/gnomad_genomes_v4"
             readoptions {}
             storageid="public_database"
             table {

diff --git a/datalake-spark3/src/test/resources/config/reference_kf.conf b/datalake-spark3/src/test/resources/config/reference_kf.conf
@@ -43,16 +43,16 @@ datalake {
         },
         {
             format=VCF
-            id="raw_gnomad_genomes_v4"
+            id="raw_gnomad_genomes_v3"
             keys=[]
             loadtype=OverWrite
             partitionby=[]
-            path="/release/4.1/vcf/genomes/gnomad.genomes.v4.1.sites.chr[^M]*.vcf.bgz"
+            path="/release/3.1/vcf/genomes/gnomad.genomes.v3.1.sites.chr[^M]*.vcf.bgz"
             readoptions {
                 flattenInfoFields="true"
                 "split_multiallelics"="true"
             }
-            storageid=gnomadv4
+            storageid=gnomad
             writeoptions {
                 "created_on_column"="created_on"
                 "is_current_column"="is_current"
@@ -63,11 +63,11 @@ datalake {
         },
         {
             format=VCF
-            id="raw_gnomad_genomes_v3"
+            id="raw_gnomad_genomes_v4"
             keys=[]
             loadtype=OverWrite
             partitionby=[]
-            path="/release/3.1/vcf/genomes/gnomad.genomes.v3.1.sites.chr[^M]*.vcf.bgz"
+            path="/release/4.1/vcf/genomes/gnomad.genomes.v4.1.sites.chr[^M]*.vcf.bgz"
             readoptions {
                 flattenInfoFields="true"
                 "split_multiallelics"="true"
@@ -885,7 +885,7 @@ datalake {
             storageid="public_database"
             table {
                 database=variant
-                name="gnomad_genomes_v3"
+                name="gnomad_genomes_v4"
             }
             view {
                 database="variant_live"
@@ -898,7 +898,7 @@ datalake {
                 "valid_from_column"="valid_from"
                 "valid_to_column"="valid_to"
             }
-        }
+        },
         {
             format=DELTA
             id="normalized_human_genes"