Use glob not walk for process execution directory search

Author: samirelanduk

nextflow/io.py

@@ -1,4 +1,5 @@
 import os
+import glob
 from datetime import datetime
 
 def get_file_text(path):
@@ -35,10 +36,7 @@ def get_process_ids_to_paths(process_ids, execution_path):
 
     process_ids_to_paths = {}
     path = os.path.join(execution_path, "work")
-    subdirectories = []
-    for root, dirs, _ in os.walk(path):
-        for directory in dirs:
-            subdirectories.append(os.path.join(root, directory))
+    subdirectories = glob.glob(os.path.join(path, "*", "*"))
     for subdirectory in subdirectories:
         sub = os.path.sep.join(subdirectory.split(os.path.sep)[-2:])
         for process_id in process_ids:

tests/unit/test_io.py

@@ -38,14 +38,13 @@ def test_can_handle_no_file(self, mock_getctime):
 
 class ProcessIdsToPathsTest(TestCase):
 
-    @patch("os.walk")
-    def test_can_get_paths(self, mock_walk):
+    @patch("glob.glob")
+    def test_can_get_paths(self, mock_glob):
         process_ids = ["ab/123456", "cd/7890123"]
-        mock_walk.return_value = (
-            ("/ex/work", ["xx", "cd"], []),
-            ("/ex/work/xx", ["yyyyyyy"], []),
-            ("/ex/work/cd", ["789012345678"], []),
+        mock_glob.return_value = (
+            "/ex/work/xx/yyyyyyy",
+            "/ex/work/cd/789012345678"
         )
         paths = get_process_ids_to_paths(process_ids, "/ex")
         self.assertEqual(paths, {"cd/7890123": os.path.join("/ex", "work", "cd/789012345678")})
-        mock_walk.assert_called_with(os.path.join("/ex", "work"))
+        mock_glob.assert_called_with(os.path.join("/ex", "work", "*", "*"))