Fix bug where the checksum of zipfiles is wrong (#930)

* Fix bug where the checksum of zipfiles is wrong

This bug is caused by incorrect length being written to the file, because Zipfile thinks the len() of the passed object is the length in bytes, but it was passing a ndarray, whose len() is the number of rows. The fix is to convert to bytes before passing to zipfile.writestr()

* add test and use a view instead of a copy

Co-authored-by: jakirkham <[email protected]>

Author: orenwatson

zarr/storage.py

@@ -1551,7 +1551,7 @@ def __getitem__(self, key):
     def __setitem__(self, key, value):
         if self.mode == 'r':
             raise ReadOnlyError()
-        value = ensure_contiguous_ndarray(value)
+        value = ensure_contiguous_ndarray(value).view("u1")
         with self.mutex:
             # writestr(key, value) writes with default permissions from
             # zipfile (600) that are too restrictive, build ZipInfo for

zarr/tests/test_storage.py

@@ -1554,6 +1554,13 @@ def test_permissions(self):
             assert perm == '0o40775'
         z.close()
 
+    def test_store_and_retrieve_ndarray(self):
+        store = ZipStore('data/store.zip')
+        x = np.array([[1, 2], [3, 4]])
+        store['foo'] = x
+        y = np.frombuffer(store['foo'], dtype=x.dtype).reshape(x.shape)
+        assert np.array_equiv(y, x)
+
 
 class TestDBMStore(StoreTests):