From 51e0c96d2b542da9420f468bc39d81dbb2684885 Mon Sep 17 00:00:00 2001
From: Cas Simons <cas.simons@populationgenomics.org.au>
Date: Thu, 17 Aug 2023 09:54:39 +1200
Subject: [PATCH] Warn if there are incomplete fastq pairs

---
 scripts/redcap_parsing_utils.py | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/scripts/redcap_parsing_utils.py b/scripts/redcap_parsing_utils.py
index d797606d7..0961f5c45 100644
--- a/scripts/redcap_parsing_utils.py
+++ b/scripts/redcap_parsing_utils.py
@@ -1,4 +1,5 @@
 # pylint: disable=unused-variable
+import sys
 from enum import Enum
 from collections import defaultdict
 from cloudpathlib import CloudPath
@@ -271,6 +272,9 @@ def find_fastq_pairs(
     # Group by sample
     read_pairs_by_sample_id = defaultdict(list)
     for pair in read_pairs.values():
+        if len(pair) != 2:
+            print(f"Warning: skipping incomplete pair: {[fq.path for fq in pair]}", file=sys.stderr)
+            continue
         read_pairs_by_sample_id[pair[0].sample_id].append(pair)
 
     return read_pairs_by_sample_id