Fail Avro extract and callset stats on bad filter name [VS-655] #8046
Conversation
There was a problem hiding this comment.
there's already a task ValidateFilterSetName in GvsExtractCallset (see https://github.com/broadinstitute/gatk/blob/ah_var_store/scripts/variantstore/wdl/GvsExtractCallset.wdl#L117); perhaps these two tasks could be combined?
Codecov Report
Additional details and impacted files@@ Coverage Diff @@
## ah_var_store #8046 +/- ##
================================================
Coverage ? 86.249%
Complexity ? 35205
================================================
Files ? 2173
Lines ? 165004
Branches ? 17791
================================================
Hits ? 142314
Misses ? 16362
Partials ? 6328 |
|
I've got a |
mcovarr
force-pushed
the
vs_655_avro_extract_warn_on_bad_filter_name
branch
from
9e961b4 to
76da384
Compare
| @@ -11,7 +11,16 @@ workflow GvsExtractAvroFilesForHail { | |||
| Int scatter_width = 10 | |||
| } | |||
|
|
|||
| call OutputPath { input: go = true } | |||
| call Utils.ValidateFilterSetName { | |||
There was a problem hiding this comment.
@rsasch gets the credit for suggesting this! 😄
| @@ -207,52 +207,11 @@ workflow GvsExtractCallset { | |||
| } | |||
| } | |||
|
|
|||
| task ValidateFilterSetName { | |||
|
|
||
| echo "project_id = ~{query_project}" > ~/.bigqueryrc | ||
|
|
||
| OUTPUT=$(bq --location=US --project_id=~{query_project} --format=csv query --use_legacy_sql=false ~{bq_labels} "SELECT filter_set_name as available_filter_set_names FROM \`~{data_project}.~{data_dataset}.filter_set_info\` GROUP BY filter_set_name") |
There was a problem hiding this comment.
Mind seeing if this works without the "--location=US" part and, if so, removing it? As is, it'll fail if we specify a dataset anywhere else. I'm 99.99% sure this call will function perfectly without that parameter.
FAILED runs with bad filter names:
Successful runs with good filter names: