GFFValidator: adding support for gzipped reference genome files

tiny/rna/collapser.py

@@ -8,21 +8,18 @@
 
 import argparse
 import builtins
-import gzip
 import os
 
 from collections import Counter
-from functools import partial
 from typing import Tuple, Iterable
 
+from tiny.rna.util import gzip_open as gz_f
+
 try:
     from _collections import _count_elements  # Load Counter's C helper function if it is available
 except ImportError:
     from collections import _count_elements   # Slower mapping[elem] = mapping.get(elem,default_val)+1
 
-# The GZIP read/write interface used by seq_counter() and seq2fasta()
-gz_f = partial(gzip.GzipFile, compresslevel=6, fileobj=None, mtime=0)
-
 
 def get_args() -> 'argparse.NameSpace':
     """Get command line arguments"""

tiny/rna/counter/validation.py

@@ -8,7 +8,7 @@
 
 from tiny.rna.counter.hts_parsing import parse_gff, ReferenceTables
 from tiny.rna.counter.features import FeatureSelector
-from tiny.rna.util import sorted_natural
+from tiny.rna.util import sorted_natural, gzip_open
 
 
 class ReportFormatter:
@@ -215,7 +215,12 @@ def chroms_shared_with_genomes(self, genome_fastas):
         for fasta in genome_fastas:
             if not os.path.isfile(fasta):
                 continue
-            with open(fasta, 'rb') as f:
+            elif fasta.endswith('.gz'):
+                file_if = gzip_open
+            else:
+                file_if = open
+
+            with file_if(fasta, 'rb') as f:
                 for line in f:
                     if line[0] == ord('>'):
                         genome_chroms.add(line[1:].strip().decode())

tiny/rna/util.py

@@ -1,6 +1,7 @@
 import argparse
 import functools
 import textwrap
+import gzip
 import time
 import os
 import re
@@ -93,4 +94,8 @@ def sorted_natural(lines, reverse=False):
 
     convert = lambda text: int(text) if text.isdigit() else text.lower()
     alphanum_key = lambda key: [convert(c) for c in re.split(r'(\d+)', key)]
-    return sorted(lines, key=alphanum_key, reverse=reverse)
+    return sorted(lines, key=alphanum_key, reverse=reverse)
+
+
+# File IO interface for reading and writing Gzip files
+gzip_open = functools.partial(gzip.GzipFile, compresslevel=6, fileobj=None, mtime=0)