Add GZip codec

docs/gzip.rst

@@ -0,0 +1,11 @@
+GZip
+====
+.. automodule:: numcodecs.gzip
+
+.. autoclass:: GZip
+
+    .. autoattribute:: codec_id
+    .. automethod:: encode
+    .. automethod:: decode
+    .. automethod:: get_config
+    .. automethod:: from_config

docs/index.rst

@@ -62,6 +62,7 @@ Contents
     lz4
     zstd
     zlib
+    gzip
     bz2
     lzma
     delta

docs/release.rst

@@ -26,6 +26,8 @@ Release notes
 
 * Add Python 3.7 (by :user:`John Kirkham <jakirkham>`; :issue:`92`).
 
+* Add codec :class:`numcodecs.gzip.GZip` to replace ``gzip`` alias for ``zlib``,
+  which was incorrect (by :user:`Jan Funke <funkey>`; :issue:`87`).
 
 .. _release_0.5.5:
 

fixture/gzip/codec.00/config.json

@@ -0,0 +1,4 @@
+{
+    "id": "gzip",
+    "level": 1
+}

fixture/gzip/codec.01/config.json

@@ -0,0 +1,4 @@
+{
+    "id": "gzip",
+    "level": -1
+}

fixture/gzip/codec.02/config.json

@@ -0,0 +1,4 @@
+{
+    "id": "gzip",
+    "level": 0
+}

fixture/gzip/codec.03/config.json

@@ -0,0 +1,4 @@
+{
+    "id": "gzip",
+    "level": 1
+}

fixture/gzip/codec.04/config.json

@@ -0,0 +1,4 @@
+{
+    "id": "gzip",
+    "level": 5
+}

fixture/gzip/codec.05/config.json

@@ -0,0 +1,4 @@
+{
+    "id": "gzip",
+    "level": 9
+}

numcodecs/__init__.py

@@ -28,7 +28,9 @@
 
 from numcodecs.zlib import Zlib
 register_codec(Zlib)
-register_codec(Zlib, 'gzip')  # alias
+
+from numcodecs.gzip import GZip
+register_codec(GZip)
 
 from numcodecs.bz2 import BZ2
 register_codec(BZ2)

numcodecs/gzip.py

@@ -0,0 +1,73 @@
+# -*- coding: utf-8 -*-
+from __future__ import absolute_import, print_function, division
+import gzip as _gzip
+import io
+
+
+import numpy as np
+
+
+from .abc import Codec
+from .compat import buffer_copy, handle_datetime, buffer_tobytes, PY2
+
+
+class GZip(Codec):
+    """Codec providing gzip compression using zlib via the Python standard library.
+
+    Parameters
+    ----------
+    level : int
+        Compression level.
+
+    """
+
+    codec_id = 'gzip'
+
+    def __init__(self, level=1):
+        self.level = level
+
+    def encode(self, buf):
+
+        # deal with lack of buffer support for datetime64 and timedelta64
+        buf = handle_datetime(buf)
+
+        if isinstance(buf, np.ndarray):
+
+            # cannot compress object array
+            if buf.dtype == object:
+                raise ValueError('cannot encode object array')
+
+            # if numpy array, can only handle C contiguous directly
+            if not buf.flags.c_contiguous:
+                buf = buf.tobytes(order='A')
+
+        if PY2:  # pragma: py3 no cover
+            # ensure bytes, PY2 cannot handle things like bytearray
+            buf = buffer_tobytes(buf)
+
+        # do compression
+        compressed = io.BytesIO()
+        with _gzip.GzipFile(fileobj=compressed,
+                            mode='wb',
+                            compresslevel=self.level) as compressor:
+            compressor.write(buf)
+        compressed = compressed.getvalue()
+
+        return compressed
+
+    # noinspection PyMethodMayBeStatic
+    def decode(self, buf, out=None):
+
+        if PY2:  # pragma: py3 no cover
+            # ensure bytes, PY2 cannot handle things like bytearray
+            buf = buffer_tobytes(buf)
+
+        # do decompression
+        buf = io.BytesIO(buf)
+        with _gzip.GzipFile(fileobj=buf, mode='rb') as decompressor:
+            decompressed = decompressor.read()
+
+        # handle destination - Python standard library zlib module does not
+        # support direct decompression into buffer, so we have to copy into
+        # out if given
+        return buffer_copy(decompressed, out)

numcodecs/tests/test_gzip.py

@@ -0,0 +1,75 @@
+# -*- coding: utf-8 -*-
+from __future__ import absolute_import, print_function, division
+import itertools
+
+
+import numpy as np
+
+
+from numcodecs.gzip import GZip
+from numcodecs.tests.common import (check_encode_decode, check_config, check_repr,
+                                    check_backwards_compatibility,
+                                    check_err_decode_object_buffer,
+                                    check_err_encode_object_buffer)
+
+
+codecs = [
+    GZip(),
+    GZip(level=-1),
+    GZip(level=0),
+    GZip(level=1),
+    GZip(level=5),
+    GZip(level=9),
+]
+
+
+# mix of dtypes: integer, float, bool, string
+# mix of shapes: 1D, 2D, 3D
+# mix of orders: C, F
+arrays = [
+    np.arange(1000, dtype='i4'),
+    np.linspace(1000, 1001, 1000, dtype='f8'),
+    np.random.normal(loc=1000, scale=1, size=(100, 10)),
+    np.random.randint(0, 2, size=1000, dtype=bool).reshape(100, 10, order='F'),
+    np.random.choice([b'a', b'bb', b'ccc'], size=1000).reshape(10, 10, 10),
+    np.random.randint(0, 2**60, size=1000, dtype='u8').view('M8[ns]'),
+    np.random.randint(0, 2**60, size=1000, dtype='u8').view('m8[ns]'),
+    np.random.randint(0, 2**25, size=1000, dtype='u8').view('M8[m]'),
+    np.random.randint(0, 2**25, size=1000, dtype='u8').view('m8[m]'),
+]
+
+
+def test_encode_decode():
+    for arr, codec in itertools.product(arrays, codecs):
+        check_encode_decode(arr, codec)
+
+
+def test_config():
+    codec = GZip(level=3)
+    check_config(codec)
+
+
+def test_repr():
+    check_repr("GZip(level=3)")
+
+
+def test_eq():
+    assert GZip() == GZip()
+    assert not GZip() != GZip()
+    assert GZip(1) == GZip(1)
+    assert GZip(1) != GZip(9)
+    assert GZip() != 'foo'
+    assert 'foo' != GZip()
+    assert not GZip() == 'foo'
+
+
+def test_backwards_compatibility():
+    check_backwards_compatibility(GZip.codec_id, arrays, codecs)
+
+
+def test_err_decode_object_buffer():
+    check_err_decode_object_buffer(GZip())
+
+
+def test_err_encode_object_buffer():
+    check_err_encode_object_buffer(GZip())

numcodecs/tests/test_zlib.py

@@ -7,7 +7,6 @@
 
 
 from numcodecs.zlib import Zlib
-from numcodecs.registry import get_codec
 from numcodecs.tests.common import (check_encode_decode, check_config, check_repr,
                                     check_backwards_compatibility,
                                     check_err_decode_object_buffer,
@@ -54,12 +53,6 @@ def test_repr():
     check_repr("Zlib(level=3)")
 
 
-def test_alias():
-    config = dict(id='gzip', level=1)
-    codec = get_codec(config)
-    assert Zlib(1) == codec
-
-
 def test_eq():
     assert Zlib() == Zlib()
     assert not Zlib() != Zlib()