some stuff to take a look at crams

src/align_genotype/scripts/diagnose_cram_manta.py

@@ -0,0 +1,43 @@
+"""
+Hail Batch job to diagnose CRAM issues causing Manta's "0 high-confidence read pairs" failure.
+
+Submits a batch job that localises the CRAM+CRAI and runs diagnose_cram_manta.sh.
+Results appear in the Hail Batch job log.
+
+Usage:
+    analysis-runner ... diagnose_cram_manta --cram_path gs://bucket/sample.cram
+"""
+
+import argparse
+
+from cpg_utils import config, hail_batch
+
+SCRIPT_PATH = '/align_genotype/src/align_genotype/scripts/diagnose_cram_manta.sh'
+
+
+def main(cram_path: str) -> None:
+    batch = hail_batch.get_batch()
+
+    job = batch.new_job('Diagnose CRAM for Manta')
+    job.image(config.config_retrieve(['workflow', 'driver_image']))
+    job.cpu(2)
+    job.memory('8Gi')
+    job.storage('100Gi')
+
+    cram = batch.read_input_group(
+        cram=cram_path,
+        crai=f'{cram_path}.crai',
+    ).cram
+
+    reference = hail_batch.fasta_res_group(batch)
+
+    job.command(f'bash {SCRIPT_PATH} {cram} {reference.base}')
+
+    batch.run(wait=False)
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description='Diagnose CRAM issues causing Manta failure')
+    parser.add_argument('--cram_path', type=str, required=True, help='GCS path to CRAM file (.crai must exist alongside)')
+    args = parser.parse_args()
+    main(cram_path=args.cram_path)