Add better test for local paths (#532)

Author: tomwhite

cubed/core/plan.py

@@ -16,7 +16,13 @@
 from cubed.runtime.pipeline import visit_nodes
 from cubed.runtime.types import ComputeEndEvent, ComputeStartEvent, CubedPipeline
 from cubed.storage.zarr import LazyZarrArray
-from cubed.utils import chunk_memory, extract_stack_summaries, join_path, memory_repr
+from cubed.utils import (
+    chunk_memory,
+    extract_stack_summaries,
+    is_local_path,
+    join_path,
+    memory_repr,
+)
 
 # A unique ID with sensible ordering, used for making directory names
 CONTEXT_ID = f"cubed-{datetime.now().strftime('%Y%m%dT%H%M%S')}-{uuid.uuid4()}"
@@ -26,7 +32,7 @@
 
 
 def delete_on_exit(context_dir: str) -> None:
-    if context_dir not in CONTEXT_DIRS and context_dir.startswith("/"):
+    if context_dir not in CONTEXT_DIRS and is_local_path(context_dir):
         atexit.register(lambda: shutil.rmtree(context_dir, ignore_errors=True))
         CONTEXT_DIRS.add(context_dir)
 

cubed/tests/test_utils.py

@@ -12,6 +12,7 @@
     block_id_to_offset,
     broadcast_trick,
     extract_stack_summaries,
+    is_local_path,
     join_path,
     map_nested,
     memory_repr,
@@ -69,6 +70,15 @@ def test_join_path():
     assert join_path("http://host/a path", "subpath") == "http://host/a%20path/subpath"
 
 
+def test_is_local_path():
+    assert is_local_path("relative_path/path")
+    assert is_local_path("/absolute_path/path")
+    assert is_local_path("file:relative_path/path")
+    assert is_local_path("file://absolute_path/path")
+    assert is_local_path("file:///absolute_path/path")
+    assert not is_local_path("s3://host/path")
+
+
 def test_memory_repr():
     assert memory_repr(0) == "0 bytes"
     assert memory_repr(1) == "1 bytes"

cubed/utils.py

@@ -74,6 +74,11 @@ def join_path(dir_url: PathType, child_path: str) -> str:
     return urlunsplit(split_parts)
 
 
+def is_local_path(path: str):
+    """Determine if a path string is for the local filesystem."""
+    return urlsplit(path).scheme in ("", "file")
+
+
 def memory_repr(num: int) -> str:
     """Convert bytes to a human-readable string in decimal form.
     1 KB is 1,000 bytes, 1 MB is 1,000,000 bytes, and so on.