From d17f823de1466c0fa2f21ae1dbdd1298a36744e6 Mon Sep 17 00:00:00 2001
From: Mitchell Robert Vollger <mrvollger@gmail.com>
Date: Tue, 11 Jan 2022 17:39:41 -0800
Subject: [PATCH] fix: added code to ignore commented lines in a bed file 
 (#474)

* Added code to ignore commented lines in a bed file and added a simple test function

* Update CHANGELOG.md

Co-authored-by: Michael Hall <michael@mbh.sh>
---
 CHANGELOG.md  |  3 +++
 src/io/bed.rs | 26 ++++++++++++++++++++++++++
 2 files changed, 29 insertions(+)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index f7ca6bcea..6d0d8cf98 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -2,6 +2,9 @@
 All notable changes to this project will be documented in this file.
 This project adheres to [Semantic Versioning](http://semver.org/).
 
+# [Unreleased](https://github.com/rust-bio/rust-bio/compare/v0.39.0...HEAD)
+- Fixed the bed reader to allow comment lines (@mrvollger).
+
 # [0.35.0] - 2021-07-05
 - Improved buffer control in Fasta and Fastq API (@natir).
 - Fixed an indexing bug in ArrayBackedIntervalTree (@wabain).
diff --git a/src/io/bed.rs b/src/io/bed.rs
index 08182eb73..f79c2ca65 100644
--- a/src/io/bed.rs
+++ b/src/io/bed.rs
@@ -54,6 +54,7 @@ impl<R: io::Read> Reader<R> {
             inner: csv::ReaderBuilder::new()
                 .delimiter(b'\t')
                 .has_headers(false)
+                .comment(Some(b'#'))
                 .from_reader(reader),
         }
     }
@@ -391,6 +392,12 @@ mod tests {
 
     const BED_FILE: &[u8] = b"1\t5\t5000\tname1\tup
 2\t3\t5005\tname2\tup
+";
+    const BED_FILE_COMMENT: &[u8] = b"\
+# this line should be ignored
+1\t5\t5000\tname1\tup
+# and this one as well
+2\t3\t5005\tname2\tup
 ";
     const BED_FILE_COMPACT: &[u8] = b"1\t5\t5000\n2\t3\t5005\n";
 
@@ -413,6 +420,25 @@ mod tests {
         }
     }
 
+    #[test]
+    fn test_reader_with_comment() {
+        let chroms = ["1", "2"];
+        let starts = [5, 3];
+        let ends = [5000, 5005];
+        let names = ["name1", "name2"];
+        let scores = ["up", "up"];
+
+        let mut reader = Reader::new(BED_FILE_COMMENT);
+        for (i, r) in reader.records().enumerate() {
+            let record = r.expect("Error reading record");
+            assert_eq!(record.chrom(), chroms[i]);
+            assert_eq!(record.start(), starts[i]);
+            assert_eq!(record.end(), ends[i]);
+            assert_eq!(record.name().expect("Error reading name"), names[i]);
+            assert_eq!(record.score().expect("Error reading score"), scores[i]);
+        }
+    }
+
     #[test]
     fn test_reader_compact() {
         let chroms = ["1", "2"];