Fixed another bug

broadinstitute · Oct 22, 2024 · 35e8aaa · 35e8aaa
1 parent 7e866e6
commit 35e8aaa
Show file tree

Hide file tree

Showing 4 changed files with 11 additions and 13 deletions.
diff --git a/scripts/variantstore/wdl/GvsExtractAvroFilesForHail.wdl b/scripts/variantstore/wdl/GvsExtractAvroFilesForHail.wdl
@@ -196,8 +196,6 @@ task ExtractFromFilterTables {
         String variants_docker
     }
 
-    String vets_score_field = 'calibration_sensitivity'
-
     parameter_meta {
         avro_sibling: "Cloud path to a file that will be the sibling to the 'avro' 'directory' under which output Avro files will be written."
     }
@@ -212,7 +210,7 @@ task ExtractFromFilterTables {
         python3 /app/run_avro_query.py --sql "
             EXPORT DATA OPTIONS(
             uri='${avro_prefix}/vets_filtering_data/vets_filtering_data_*.avro', format='AVRO', compression='SNAPPY') AS
-            SELECT location, type as model, ref, alt, ~{vets_score_field}, yng_status
+            SELECT location, type as model, ref, alt, calibration_sensitivity, yng_status
             FROM \`~{project_id}.~{dataset_name}.filter_set_info\`
             WHERE filter_set_name = '~{filter_set_name}'
             ORDER BY location

diff --git a/scripts/variantstore/wdl/GvsUtils.wdl b/scripts/variantstore/wdl/GvsUtils.wdl
@@ -72,7 +72,7 @@ task GetToolVersions {
     # GVS generally uses the smallest `alpine` version of the Google Cloud SDK as it suffices for most tasks, but
     # there are a handlful of tasks that require the larger GNU libc-based `slim`.
     String cloud_sdk_slim_docker = "gcr.io/google.com/cloudsdktool/cloud-sdk:435.0.0-slim"
-    String variants_docker = "us-central1-docker.pkg.dev/broad-dsde-methods/gvs/variants:2024-10-21-alpine-d931b2311c9e"
+    String variants_docker = "us-central1-docker.pkg.dev/broad-dsde-methods/gvs/variants:2024-10-22-alpine-e7443149b8db"
     String variants_nirvana_docker = "us.gcr.io/broad-dsde-methods/variantstore:nirvana_2022_10_19"
     String gatk_docker = "us-central1-docker.pkg.dev/broad-dsde-methods/gvs/gatk:2024_10_10-gatkbase-1cd1f9652cb9"
     String real_time_genomics_docker = "docker.io/realtimegenomics/rtg-tools:latest"

diff --git a/scripts/variantstore/wdl/extract/import_gvs.py b/scripts/variantstore/wdl/extract/import_gvs.py
@@ -168,7 +168,7 @@ def convert_array_with_id_keys_to_dense_array(arr, ids, drop=[]):
         return hl.rbind(sdict, lambda sdict: ids.map(lambda x: sdict.get(x)))
 
     site_path = os.path.join(tmp_dir, 'site_filters.ht')
-    vets_path = os.path.join(tmp_dir, 'vets.ht')
+    vets_filter_path = os.path.join(tmp_dir, 'vets_filters.ht')
 
     if intermediate_resume_point > 0:
         info('import_gvs: skipping site and VETS filter import')
@@ -183,12 +183,12 @@ def convert_array_with_id_keys_to_dense_array(arr, ids, drop=[]):
         site.write(site_path, overwrite=True)
 
         info('import_gvs: Importing and writing VETS filter data to temporary storage')
-        vets = hl.import_avro(vets_filtering_data)
-        vets = vets.transmute(
-            locus=translate_locus(vets.location)
+        vets_filter = hl.import_avro(vets_filtering_data)
+        vets_filter = vets_filter.transmute(
+            locus=translate_locus(vets_filter.location)
         )
-        vets = vets.key_by('locus')
-        vets.write(vets_path, overwrite=True)
+        vets_filter = vets_filter.key_by('locus')
+        vets_filter.write(vets_filter_path, overwrite=True)
 
     n_samples = 0
 
@@ -329,12 +329,12 @@ def convert_array_with_id_keys_to_dense_array(arr, ids, drop=[]):
 
         # read site and vets data with same intervals for efficient joins
         site = hl.read_table(site_path, _intervals=target_final_intervals)
-        vets = hl.read_table(vets_path, _intervals=target_final_intervals)
+        vets_filter = hl.read_table(vets_filter_path, _intervals=target_final_intervals)
 
         vd = vd.annotate_rows(filters=hl.coalesce(site[vd.locus].filters, hl.empty_set(hl.tstr)))
 
         # vets ref/alt come in normalized individually, so need to renormalize to the dataset ref allele
-        vd = vd.annotate_rows(as_vets = hl.dict(vets.index(vd.locus, all_matches=True)
+        vd = vd.annotate_rows(as_vets = hl.dict(vets_filter.index(vd.locus, all_matches=True)
                                                 .map(lambda record: (record.alt + vd.alleles[0][hl.len(record.ref):], record.drop('ref', 'alt')))))
 
         vd = vd.annotate_globals(truth_sensitivity_snp_threshold=truth_sensitivity_snp_threshold,

diff --git a/scripts/variantstore/wdl/test/GvsQuickstartIntegration.wdl b/scripts/variantstore/wdl/test/GvsQuickstartIntegration.wdl
@@ -5,7 +5,7 @@ import "GvsQuickstartHailIntegration.wdl" as QuickstartHailIntegration
 import "../GvsJointVariantCalling.wdl" as JointVariantCalling
 import "../GvsUtils.wdl" as Utils
 
-# comment that is still here!
+# comment that is still here!!
 
 workflow GvsQuickstartIntegration {
     input {