Dev (#931)

* Caid refactor (#927)

* CAID refactor

* lint

* lint

* remove mock

* string defaults

* format

* ruff

* add mock values

* missed a test

* ruff

* Improve allele type validation (#925)

* Improve allele type validation

* ruff

* cleanup

Author: bpblanken

v03_pipeline/lib/misc/validation.py

@@ -69,7 +69,11 @@ def validate_allele_type(
         collected_alleles = sorted(
             [tuple(x) for x in ht.aggregate(hl.agg.collect_as_set(ht.alleles))],
         )
-        msg = f'Alleles with invalid AlleleType are present in the callset: {collected_alleles}'
+        # Handle case where all invalid alleles are NON_REF, indicating a gvcf:
+        if all('<NON_REF>' in alleles for alleles in collected_alleles):
+            msg = 'Alleles with invalid allele <NON_REF> are present in the callset.  This appears to be a GVCF containing records for sites with no variants.'
+            raise SeqrValidationError(msg)
+        msg = f'Alleles with invalid AlleleType are present in the callset: {collected_alleles[:10]}'
         raise SeqrValidationError(msg)
 
 
@@ -85,7 +89,7 @@ def validate_no_duplicate_variants(
     ht = ht.select()
     if ht.count() > 0:
         variant_format = dataset_type.table_key_format_fn(reference_genome)
-        msg = f'Variants are present multiple times in the callset: {[variant_format(v) for v in ht.collect()]}'
+        msg = f'Variants are present multiple times in the callset: {[variant_format(v) for v in ht.collect()][:10]}'
         raise SeqrValidationError(msg)
 
 

v03_pipeline/lib/misc/validation_test.py

@@ -81,6 +81,36 @@ def test_validate_allele_type(self) -> None:
             DatasetType.SNV_INDEL,
         )
 
+        mt = hl.MatrixTable.from_parts(
+            rows={
+                'locus': [
+                    hl.Locus(
+                        contig='chr1',
+                        position=1,
+                        reference_genome='GRCh38',
+                    ),
+                    hl.Locus(
+                        contig='chr1',
+                        position=2,
+                        reference_genome='GRCh38',
+                    ),
+                ],
+                'alleles': [
+                    ['C', '<NON_REF>'],
+                    ['A', '<NON_REF>'],
+                ],
+            },
+            cols={'s': ['sample_1']},
+            entries={'HL': [[0.0], [0.0]]},
+        ).key_rows_by('locus', 'alleles')
+        self.assertRaisesRegex(
+            SeqrValidationError,
+            'Alleles with invalid allele <NON_REF> are present in the callset.  This appears to be a GVCF containing records for sites with no variants.',
+            validate_allele_type,
+            mt,
+            DatasetType.SNV_INDEL,
+        )
+
     @patch('v03_pipeline.lib.misc.validation.Env')
     def test_validate_imputed_sex_ploidy(self, mock_env: Mock) -> None:
         mock_env.CHECK_SEX_AND_RELATEDNESS = True

v03_pipeline/lib/tasks/validate_callset_test.py

@@ -83,7 +83,7 @@ def test_validate_callset_multiple_exceptions(
                 json.load(f),
                 {
                     'error_messages': [
-                        "Alleles with invalid AlleleType are present in the callset: [('G', '<NON_REF>')]",
+                        'Alleles with invalid allele <NON_REF> are present in the callset.  This appears to be a GVCF containing records for sites with no variants.',
                         "Variants are present multiple times in the callset: ['1-902088-G-A']",
                         'Missing the following expected contigs:chr10, chr11, chr12, chr13, chr14, chr15, chr16, chr17, chr18, chr19, chr2, chr20, chr21, chr22, chr3, chr4, chr5, chr6, chr7, chr8, chr9, chrX',
                         'Sample type validation error: dataset sample-type is specified as WES but appears to be WGS because it contains many common non-coding variants',