Merge pull request #69 from alimanfoo/issue_65

Group supports item deletion; resolves #65

Author: alimanfoo

docs/release.rst

@@ -1,6 +1,8 @@
 Release notes
 =============
 
+* Group objects now support member deletion via ``del`` statement
+  ('#65 <https://github.com/alimanfoo/zarr/issues/65>'_)
 * Added :class:`zarr.storage.TempStore` class for convenience to provide
   storage via a temporary directory
   (`#59 <https://github.com/alimanfoo/zarr/issues/59>`_)

zarr/creation.py

@@ -289,20 +289,15 @@ def array(data, **kwargs):
         data = np.asanyarray(data)
 
     # setup dtype
-    dtype = kwargs.pop('dtype', None)
-    if dtype is None:
-        dtype = data.dtype
+    kwargs.setdefault('dtype', data.dtype)
 
-    # setup shape
-    shape = data.shape
-
-    # setup chunks
-    chunks = kwargs.pop('chunks', None)
-    if chunks is None:
-        _, chunks = _get_shape_chunks(data)
+    # setup shape and chunks
+    shape, chunks = _get_shape_chunks(data)
+    kwargs['shape'] = data.shape
+    kwargs.setdefault('chunks', chunks)
 
     # instantiate array
-    z = create(shape=shape, chunks=chunks, dtype=dtype, **kwargs)
+    z = create(**kwargs)
 
     # fill with data
     z[:] = data

zarr/hierarchy.py

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 from __future__ import absolute_import, print_function, division
-from collections import Mapping
+from collections import MutableMapping
 
 
 import numpy as np
@@ -9,7 +9,7 @@
 from zarr.attrs import Attributes
 from zarr.core import Array
 from zarr.storage import contains_array, contains_group, init_group, \
-    DictStore, DirectoryStore, group_meta_key, attrs_key, listdir
+    DictStore, DirectoryStore, group_meta_key, attrs_key, listdir, rmdir
 from zarr.creation import array, create, empty, zeros, ones, full, \
     empty_like, zeros_like, ones_like, full_like
 from zarr.util import normalize_storage_path, normalize_shape
@@ -18,7 +18,7 @@
 from zarr.meta import decode_group_metadata
 
 
-class Group(Mapping):
+class Group(MutableMapping):
     """Instantiate a group from an initialized store.
 
     Parameters
@@ -303,6 +303,20 @@ def __getitem__(self, item):
         else:
             raise KeyError(item)
 
+    def __setitem__(self, item, value):
+        self.array(item, value, overwrite=True)
+
+    def __delitem__(self, item):
+        return self._write_op(self._delitem_nosync, item)
+
+    def _delitem_nosync(self, item):
+        path = self._item_path(item)
+        if contains_array(self._store, path) or \
+                contains_group(self._store, path):
+            rmdir(self._store, path)
+        else:
+            raise KeyError(item)
+
     def group_keys(self):
         """Return an iterator over member names for groups only.
 
@@ -494,10 +508,7 @@ def require_groups(self, *names):
         """Convenience method to require multiple groups in a single call."""
         return tuple(self.require_group(name) for name in names)
 
-    def create_dataset(self, name, data=None, shape=None, chunks=None,
-                       dtype=None, compressor='default', fill_value=0,
-                       order='C', synchronizer=None, filters=None,
-                       overwrite=False, cache_metadata=True, **kwargs):
+    def create_dataset(self, name, **kwargs):
         """Create an array.
 
         Parameters
@@ -550,43 +561,23 @@ def create_dataset(self, name, data=None, shape=None, chunks=None,
 
         """  # flake8: noqa
 
-        return self._write_op(self._create_dataset_nosync, name, data=data,
-                              shape=shape, chunks=chunks, dtype=dtype,
-                              compressor=compressor, fill_value=fill_value,
-                              order=order, synchronizer=synchronizer,
-                              filters=filters, overwrite=overwrite,
-                              cache_metadata=cache_metadata, **kwargs)
+        return self._write_op(self._create_dataset_nosync, name, **kwargs)
 
-    def _create_dataset_nosync(self, name, data=None, shape=None, chunks=None,
-                               dtype=None, compressor='default',
-                               fill_value=0, order='C', synchronizer=None,
-                               filters=None, overwrite=False,
-                               cache_metadata=True, **kwargs):
+    def _create_dataset_nosync(self, name, data=None, **kwargs):
 
         path = self._item_path(name)
 
         # determine synchronizer
-        if synchronizer is None:
-            synchronizer = self._synchronizer
+        kwargs.setdefault('synchronizer', self._synchronizer)
 
         # create array
-        if data is not None:
-            a = array(data, chunks=chunks, dtype=dtype,
-                      compressor=compressor, fill_value=fill_value,
-                      order=order, synchronizer=synchronizer,
-                      store=self._store, path=path,
-                      chunk_store=self._chunk_store, filters=filters,
-                      overwrite=overwrite, cache_metadata=cache_metadata,
-                      **kwargs)
+        if data is None:
+            a = create(store=self._store, path=path,
+                       chunk_store=self._chunk_store, **kwargs)
 
         else:
-            a = create(shape=shape, chunks=chunks, dtype=dtype,
-                       compressor=compressor, fill_value=fill_value,
-                       order=order, synchronizer=synchronizer,
-                       store=self._store, path=path,
-                       chunk_store=self._chunk_store, filters=filters,
-                       overwrite=overwrite, cache_metadata=cache_metadata,
-                       **kwargs)
+            a = array(data, store=self._store, path=path,
+                      chunk_store=self._chunk_store, **kwargs)
 
         return a
 

zarr/storage.py

@@ -891,8 +891,8 @@ def getsize(self, path=None):
                 return info.compress_size
             except KeyError:
                 err_path_not_found(path)
-            else:
-                return 0
+        else:
+            return 0
 
 
 def migrate_1to2(store):

zarr/tests/test_creation.py

@@ -72,6 +72,8 @@ def test_array():
     eq(z.dtype, z2.dtype)
     assert_array_equal(z[:], z2[:])
 
+    # with chunky array-likes
+
     b = np.arange(1000).reshape(100, 10)
     c = MockBcolzArray(b, 10)
     z3 = array(c)

zarr/tests/test_hierarchy.py

@@ -507,8 +507,33 @@ def test_group_repr(self):
 
     def test_setitem(self):
         g = self.create_group()
-        with assert_raises(TypeError):
-            g['foo'] = 'bar'
+        try:
+            data = np.arange(100)
+            g['foo'] = data
+            assert_array_equal(data, g['foo'])
+            data = np.arange(200)
+            g['foo'] = data
+            assert_array_equal(data, g['foo'])
+        except NotImplementedError:
+            pass
+
+    def test_delitem(self):
+        g = self.create_group()
+        g.create_group('foo')
+        g.create_dataset('bar/baz', shape=100, chunks=10)
+        assert 'foo' in g
+        assert 'bar' in g
+        assert 'bar/baz' in g
+        try:
+            del g['bar']
+            with assert_raises(KeyError):
+                del g['xxx']
+        except NotImplementedError:
+            pass
+        else:
+            assert 'foo' in g
+            assert 'bar' not in g
+            assert 'bar/baz' not in g
 
     def test_array_creation(self):
         grp = self.create_group()