From 932e872af982c6358d153d81bf308821f8ebd8ff Mon Sep 17 00:00:00 2001
From: Sergey Koren <sergek@umd.edu>
Date: Mon, 13 May 2024 11:59:35 -0400
Subject: [PATCH] ensure we fail when conversion of hic pairs fails, don't
 track mis-sorted reads to save memory

---
 src/Snakefiles/8-hicPipeline.sm |  4 ++--
 src/scripts/parse_sam_pairs.py  | 10 ----------
 2 files changed, 2 insertions(+), 12 deletions(-)

diff --git a/src/Snakefiles/8-hicPipeline.sm b/src/Snakefiles/8-hicPipeline.sm
index 0471ae8b..0f5e2f98 100644
--- a/src/Snakefiles/8-hicPipeline.sm
+++ b/src/Snakefiles/8-hicPipeline.sm
@@ -386,8 +386,8 @@ rule transformBWA:
 cd 8-hicPipeline
 
 cat > ./transform_bwa.sh <<EOF
-#!/bin/sh
-set -e
+#!/usr/bin/env bash
+set -e -o pipefail
 {params.SAMTOOLS} view -q 1 ../{input.bwa_mapping} | {PYTHON} {VERKKO}/scripts/parse_sam_pairs.py  > ../{output.byread_mapping}
 EOF
 
diff --git a/src/scripts/parse_sam_pairs.py b/src/scripts/parse_sam_pairs.py
index 30e6ec24..fdb2f04a 100755
--- a/src/scripts/parse_sam_pairs.py
+++ b/src/scripts/parse_sam_pairs.py
@@ -33,25 +33,15 @@ def print_results(names):
 
 name = ""
 names = [ ]
-seen = {}
-out_of_order = 0
 
 for line in input_stream:
    line=line.split()
    if name == "":
       name = line[0]
    if name != line[0]:
-      seen[name] = 1
       print_results(names)
       name = line[0]
       names = [ ]
-   if name in seen:
-      print("Warning: read %s already seen but encountered it again, please confirm your bam file is sorted by read."%(name), file=sys.stderr)
-      out_of_order += 1
    names.append("%s\t%s"%(line[0], line[2]))
 
-if out_of_order > 1000:
-   print("Error: encountered too many unsorted reads (%d), exiting. Please confirm the input bam is sorted by read."%(out_of_order), file=sys.stderr)
-   sys.exit(1)
-
 print_results(names)