remove_file_extensions is more robust to various fov naming schemes (…

…#577)

* remove_file_extensions is more robust to fov naming schemes

* added feather to extension_types

* typo

Co-authored-by: Noah F. Greenwald <noahfgreenwald@gmail.com>

ark/utils/io_utils.py

@@ -1,3 +1,4 @@
+from argparse import Namespace
 import os
 import pathlib
 import warnings
@@ -112,16 +113,28 @@ def remove_file_extensions(files):
         return
 
     # remove the file extension
-    names = [os.path.splitext(name)[0] for name in files]
-
-    # identify names with '.' in them: these may not be processed correctly
-    bad_names = [name for name in names if '.' in name]
+    names = [os.path.splitext(name) for name in files]
+    names_corrected = []
+    extension_types = ["tiff", "tif", "png", "jpg", "jpeg", "tar", "gz", "csv", "feather"]
+    for name in names:
+        # We want everything after the "." for the extension
+        ext = name[-1][1:]
+        if (ext in extension_types) or (len(ext) == 0):
+            # If it is one of the extension types, only keep the filename.
+            # Or there is no extension and the names are similar to ["fov1", "fov2", "fov3", ...]
+            names_corrected.append(name[:-1][0])
+        else:
+            # If `ext` not one of the specified file types, keep the value after the "."
+            names_corrected.append(name[:-1][0] + "." + name[-1][1])
+
+    # identify names with '.' in them: these may not be processed correctly.
+    bad_names = [name for name in names_corrected if '.' in name]
     if len(bad_names) > 0:
         warnings.warn(f"These files still have \".\" in them after file extension removal: "
                       f"{','.join(bad_names)}, "
                       f"please double check that these are the correct names")
 
-    return names
+    return names_corrected
 
 
 def extract_delimited_names(names, delimiter='_', delimiter_optional=True):

ark/utils/io_utils_test.py

@@ -139,6 +139,7 @@ def test_list_files():
 def test_remove_file_extensions():
     # test a mixture of file paths and extensions
     files = ['fov1.tiff', 'fov2.tif', 'fov3.png', 'fov4.jpg']
+    files2 = ['fov.1.tiff', 'fov.2.tiff', 'fov.3.png', 'fov.4']
 
     assert iou.remove_file_extensions(None) is None
     assert iou.remove_file_extensions([]) == []
@@ -153,6 +154,10 @@ def test_remove_file_extensions():
         new_files = iou.remove_file_extensions(['fov5.tar.gz', 'fov6.sample.csv'])
         assert new_files == ['fov5.tar', 'fov6.sample']
 
+    with pytest.warns(UserWarning):
+        new_files = iou.remove_file_extensions(files2)
+        assert new_files == ['fov.1', 'fov.2', 'fov.3', 'fov.4']
+
 
 def test_extract_delimited_names():
     filenames = [