Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Enforce dtype=object for incompatible numpy array conversion #417

Merged
merged 2 commits into from
Jan 13, 2023
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
5 changes: 4 additions & 1 deletion numcodecs/json.py
Original file line number Diff line number Diff line change
Expand Up @@ -54,7 +54,10 @@ def __init__(self, encoding='utf-8', skipkeys=False, ensure_ascii=True,
self._decoder = _json.JSONDecoder(**self._decoder_config)

def encode(self, buf):
buf = np.asarray(buf)
try:
buf = np.asarray(buf)
except ValueError:
buf = np.asarray(buf, dtype=object)
items = buf.tolist()
items.extend((buf.dtype.str, buf.shape))
return self._encoder.encode(items).encode(self._text_encoding)
Expand Down
5 changes: 4 additions & 1 deletion numcodecs/msgpacks.py
Original file line number Diff line number Diff line change
Expand Up @@ -52,7 +52,10 @@ def __init__(self, use_single_float=False, use_bin_type=True, raw=False):
self.raw = raw

def encode(self, buf):
buf = np.asarray(buf)
try:
buf = np.asarray(buf)
except ValueError:
buf = np.asarray(buf, dtype=object)
martindurant marked this conversation as resolved.
Show resolved Hide resolved
items = buf.tolist()
items.extend((buf.dtype.str, buf.shape))
return msgpack.packb(items, use_bin_type=self.use_bin_type,
Expand Down
36 changes: 19 additions & 17 deletions numcodecs/tests/test_json.py
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@


import numpy as np

import pytest

from numcodecs.json import JSON
from numcodecs.tests.common import (check_config, check_repr, check_encode_decode_array,
Expand Down Expand Up @@ -53,21 +53,23 @@ def test_backwards_compatibility():
check_backwards_compatibility(JSON.codec_id, arrays, codecs)


def test_non_numpy_inputs():
@pytest.mark.parametrize(
"input_data, dtype",
[
([0, 1], None),
([[0, 1], [2, 3]], None),
([[0], [1], [2, 3]], object),
([[[0, 0]], [[1, 1]], [[2, 3]]], None),
(["1"], None),
(["11", "11"], None),
(["11", "1", "1"], None),
([{}], None),
([{"key": "value"}, ["list", "of", "strings"]], object),
]
)
def test_non_numpy_inputs(input_data, dtype):
# numpy will infer a range of different shapes and dtypes for these inputs.
# Make sure that round-tripping through encode preserves this.
data = [
[0, 1],
[[0, 1], [2, 3]],
[[0], [1], [2, 3]],
[[[0, 0]], [[1, 1]], [[2, 3]]],
["1"],
["11", "11"],
["11", "1", "1"],
[{}],
[{"key": "value"}, ["list", "of", "strings"]],
]
for input_data in data:
for codec in codecs:
output_data = codec.decode(codec.encode(input_data))
assert np.array_equal(np.array(input_data), output_data)
for codec in codecs:
output_data = codec.decode(codec.encode(input_data))
assert np.array_equal(np.array(input_data, dtype=dtype), output_data)
45 changes: 24 additions & 21 deletions numcodecs/tests/test_msgpacks.py
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,7 @@


import numpy as np
import pytest


try:
Expand Down Expand Up @@ -52,30 +53,32 @@ def test_backwards_compatibility():
check_backwards_compatibility(codec.codec_id, arrays, [codec])


def test_non_numpy_inputs():
@pytest.mark.parametrize(
"input_data, dtype",
[
([0, 1], None),
([[0, 1], [2, 3]], None),
([[0], [1], [2, 3]], object),
([[[0, 0]], [[1, 1]], [[2, 3]]], None),
(["1"], None),
(["11", "11"], None),
(["11", "1", "1"], None),
([{}], None),
([{"key": "value"}, ["list", "of", "strings"]], object),
([b"1"], None),
([b"11", b"11"], None),
([b"11", b"1", b"1"], None),
([{b"key": b"value"}, [b"list", b"of", b"strings"]], object),
]
)
def test_non_numpy_inputs(input_data, dtype):
codec = MsgPack()
# numpy will infer a range of different shapes and dtypes for these inputs.
# Make sure that round-tripping through encode preserves this.
data = [
[0, 1],
[[0, 1], [2, 3]],
[[0], [1], [2, 3]],
[[[0, 0]], [[1, 1]], [[2, 3]]],
["1"],
["11", "11"],
["11", "1", "1"],
[{}],
[{"key": "value"}, ["list", "of", "strings"]],
[b"1"],
[b"11", b"11"],
[b"11", b"1", b"1"],
[{b"key": b"value"}, [b"list", b"of", b"strings"]],
]
for input_data in data:
actual = codec.decode(codec.encode(input_data))
expect = np.array(input_data)
assert expect.shape == actual.shape
assert np.array_equal(expect, actual)
actual = codec.decode(codec.encode(input_data))
expect = np.array(input_data, dtype=dtype)
assert expect.shape == actual.shape
assert np.array_equal(expect, actual)


def test_encode_decode_shape_dtype_preserved():
Expand Down