[ah_var_store] Retry with more memory VCF extract [VS-1351] (#8812)

.dockstore.yml

@@ -164,7 +164,7 @@ workflows:
          branches:
              - master
              - ah_var_store
-             - vs_1334_vcf_max_alt_alleles
+             - vs_1351_retry_with_more_memory_vcf_extract
          tags:
              - /.*/
    - name: GvsImportGenomes

scripts/variantstore/docs/aou/AOU_DELIVERABLES.md

@@ -93,7 +93,7 @@
         - This workflow does not use the Terra Data Entity Model to run, so be sure to select the `Run workflow with inputs defined by file paths` workflow submission option.
     - Specify the same `call_set_identifier`, `dataset_name`, `project_id`, `extract_table_prefix`, and `interval_list` that were used in the `GvsPrepareRangesCallset` run documented above.
     - Specify the `interval_weights_bed` appropriate for the PGEN / VCF extraction run you are performing. `gs://gvs_quickstart_storage/weights/gvs_full_vet_weights_1kb_padded_orig.bed` is the interval weights BED used for Quickstart.
-    - For `GvsExtractCallsetPgenMerged` only, select the workflow option "Retry with more memory" and choose a "Memory retry factor" of 1.5
+    - For both `GvsExtractCallset` and `GvsExtractCallsetPgenMerged`, select the workflow option "Retry with more memory" and choose a "Memory retry factor" of 1.5
     - For `GvsExtractCallset`, make sure to specify the appropriate `maximum_alternate_alleles` value (currently 100).
     - These workflows do not use the Terra Data Entity Model to run, so be sure to select the `Run workflow with inputs defined by file paths` workflow submission option.
 1. `GvsCalculatePrecisionAndSensitivity` workflow

scripts/variantstore/wdl/GvsExtractCallset.wdl

@@ -371,7 +371,18 @@ task ExtractTask {
         --filter-set-name ~{filter_set_name}'
     fi
 
-    gatk --java-options "-Xmx~{memory_gib - 3}g" \
+    # This tool may get invoked with "Retry with more memory" with a different amount of memory than specified in
+    # the input `memory_gib`, so use the memory-related environment variables rather than the `memory_gib` input.
+    # https://support.terra.bio/hc/en-us/articles/4403215299355-Out-of-Memory-Retry
+    if [[ ${MEM_UNIT} == "GB" ]]
+    then
+        memory_mb=$(python3 -c "from math import floor; print(floor((${MEM_SIZE} - 3) * 1000))")
+    else
+        echo "Unexpected memory unit: ${MEM_UNIT}" 1>&2
+        exit 1
+    fi
+
+    gatk --java-options "-Xmx${memory_mb}m" \
       ExtractCohortToVcf \
         --vet-ranges-extract-fq-table ~{fq_ranges_cohort_vet_extract_table} \
         ~{"--vet-ranges-extract-table-version " + vet_extract_table_version} \