Replaced np.bool refs; fix CI failure #34835

WillAyd · 2020-06-16T23:48:06Z

I mostly just replace np.bool with np.bool_ though there are some other cases where I just used bool. I don't know if it matters...so happy to conform one way or another

For now just seeing if this fixes CI

Closes #34848.

jbrockmendel · 2020-06-17T02:05:40Z

Looks like also need np.complex?

TomAugspurger · 2020-06-17T13:09:11Z

Pushed a batch of np.object -> object. There are still a few more failures.

TomAugspurger · 2020-06-17T13:56:48Z

The last run (https://travis-ci.org/github/pandas-dev/pandas/jobs/699299640#L3736) passed, but there were still many warnings. Because we don't run that job with -Werror.

Anyway, hopefully these are all caught now.

pandas/_libs/hashtable_func_helper.pxi.in

TomAugspurger · 2020-06-17T13:59:21Z

pandas/tests/frame/methods/test_to_records.py

@@ -163,7 +163,7 @@ def test_to_records_with_categorical(self):
            ),
            # Pass in a type instance.
            (
-                dict(column_dtypes=np.unicode),
+                dict(column_dtypes=str),


Since we only support python3 str works here.

TomAugspurger · 2020-06-17T13:59:58Z

pandas/tests/generic/test_finalize.py

-        "week",
-        "weekofyear",


This fixes warnings coming from pandas. Unrelated to the numpy changes.

jorisvandenbossche · 2020-06-17T14:09:28Z

pandas/core/dtypes/cast.py

@@ -1483,7 +1483,7 @@ def find_common_type(types: List[DtypeObj]) -> DtypeObj:
    if has_bools:
        for t in types:
            if is_integer_dtype(t) or is_float_dtype(t) or is_complex_dtype(t):
-                return np.object
+                return object


Suggested change

return object

return np.dtype("object")

(to ensure this function actually returns a dtype object, as its type annotation indicates)

jorisvandenbossche · 2020-06-17T14:11:03Z

pandas/core/util/hashing.py

@@ -264,7 +264,7 @@ def hash_array(

    # First, turn whatever array this is into unsigned 64-bit ints, if we can
    # manage it.
-    elif isinstance(dtype, np.bool):
+    elif isinstance(dtype, bool):


I am not sure if this can ever be correct:

In [18]: isinstance(np.dtype(bool), bool) Out[18]: False

unless dtype is not actually a dtype ...

jreback · 2020-06-17T14:13:09Z

after this would be nice to have a lint rule to catch these if possible (as it would fail much sooner than the 3.9 CI)

TomAugspurger · 2020-06-17T14:17:26Z

Opened #34852 to fail on warnings for that build. I'd like to get CI passing first for other PRs though.

TomAugspurger · 2020-06-17T14:24:49Z

Good to merge when CI passes?

TomAugspurger · 2020-06-17T14:33:14Z

Has anyone seen this failure on npdev before? Restarting that build.

=================================== FAILURES ===================================
________________ TestDataFrameSortIndex.test_sort_index_inplace ________________
[gw1] linux -- Python 3.7.7 /home/vsts/miniconda3/envs/pandas-dev/bin/python

self = <pandas.tests.frame.methods.test_sort_index.TestDataFrameSortIndex object at 0x7f566041fa50>

    def test_sort_index_inplace(self):
        frame = DataFrame(
            np.random.randn(4, 4), index=[1, 2, 3, 4], columns=["A", "B", "C", "D"]
        )
    
        # axis=0
        unordered = frame.loc[[3, 2, 4, 1]]
        a_id = id(unordered["A"])
        df = unordered.copy()
        df.sort_index(inplace=True)
        expected = frame
        tm.assert_frame_equal(df, expected)
>       assert a_id != id(df["A"])
E       assert 140008958850064 != 140008958850064
E        +  where 140008958850064 = id(1   -1.103603\n2   -0.373049\n3   -0.418311\n4   -0.079376\nName: A, dtype: float64)

pandas/tests/frame/methods/test_sort_index.py:224: AssertionError

TomAugspurger · 2020-06-17T14:50:56Z

The 3.9 build passed, with a few warnings I can't track down yet

=============================== warnings summary ===============================

pandas/core/arrays/boolean.py:735

pandas/core/arrays/boolean.py:735

pandas/core/arrays/boolean.py:735

  /home/travis/build/pandas-dev/pandas/pandas/core/arrays/boolean.py:735: DeprecationWarning: In future, it will be an error for 'np.bool_' scalars to be interpreted as an index

    result = op(self._data, other)

pandas/_testing.py:699

pandas/_testing.py:699

pandas/_testing.py:699

pandas/_testing.py:699

pandas/_testing.py:699

pandas/_testing.py:699

pandas/_testing.py:699

pandas/_testing.py:699

  /home/travis/build/pandas-dev/pandas/pandas/_testing.py:699: DeprecationWarning: elementwise comparison failed; this will raise an error in the future.

    diff = np.sum((left.values != right.values).astype(int)) * 100.0 / len(left)

pandas/io/json/_json.py:158: 15 tests with warnings

  /home/travis/build/pandas-dev/pandas/pandas/io/json/_json.py:158: DeprecationWarning: an integer is required (got type float).  Implicit conversion to integers using __int__ is deprecated, and may be removed in a future version of Python.

    return dumps(

pandas/tests/io/json/test_pandas.py:1176

pandas/tests/io/json/test_pandas.py:1176

pandas/tests/io/json/test_pandas.py:1176

  /home/travis/build/pandas-dev/pandas/pandas/tests/io/json/test_pandas.py:1176: DeprecationWarning: an integer is required (got type float).  Implicit conversion to integers using __int__ is deprecated, and may be removed in a future version of Python.

    assert dumps(ts, iso_dates=True) == exp

pandas/tests/io/json/test_pandas.py:1178

pandas/tests/io/json/test_pandas.py:1178

pandas/tests/io/json/test_pandas.py:1178

  /home/travis/build/pandas-dev/pandas/pandas/tests/io/json/test_pandas.py:1178: DeprecationWarning: an integer is required (got type float).  Implicit conversion to integers using __int__ is deprecated, and may be removed in a future version of Python.

    assert dumps(dt, iso_dates=True) == exp

pandas/tests/io/json/test_pandas.py:1202

pandas/tests/io/json/test_pandas.py:1202

pandas/tests/io/json/test_pandas.py:1202

pandas/tests/io/json/test_pandas.py:1202

pandas/tests/io/json/test_pandas.py:1202

pandas/tests/io/json/test_pandas.py:1202

  /home/travis/build/pandas-dev/pandas/pandas/tests/io/json/test_pandas.py:1202: DeprecationWarning: an integer is required (got type float).  Implicit conversion to integers using __int__ is deprecated, and may be removed in a future version of Python.

    result = dumps(df, iso_dates=True)

I'll handle those in #34852

TomAugspurger · 2020-06-17T15:36:59Z

All green @WillAyd @jreback.

WillAyd · 2020-06-17T15:42:43Z

lgtm @jreback . thanks Tom for pushing this through

WillAyd · 2020-06-17T15:43:09Z

lgtm @jreback . thanks Tom for pushing this through

eric-wieser · 2020-06-30T11:44:38Z

xref numpy/numpy#14882

Replaced np.bool refs

0795f80

simonjayhawkins mentioned this pull request Jun 17, 2020

DOC: 1.0.5 release date #34845

Merged

jorisvandenbossche mentioned this pull request Jun 17, 2020

CI: Failing np.bool on numpy master #34848

Closed

TomAugspurger added 3 commits June 17, 2020 07:47

np.bool is an alias

c42778a

update complex

131a894

np.object -> object

2c2517b

jreback added the CI Continuous Integration label Jun 17, 2020

jreback added this to the 1.1 milestone Jun 17, 2020

float

aaf14fa

TomAugspurger marked this pull request as ready for review June 17, 2020 13:13

TomAugspurger added 2 commits June 17, 2020 08:33

sparse

6870d19

more

f29ef6e

TomAugspurger reviewed Jun 17, 2020

View reviewed changes

pandas/_libs/hashtable_func_helper.pxi.in Outdated Show resolved Hide resolved

TomAugspurger reviewed Jun 17, 2020

View reviewed changes

cleanup

af1b945

jorisvandenbossche reviewed Jun 17, 2020

View reviewed changes

TomAugspurger mentioned this pull request Jun 17, 2020

CI: Fail on warning for 3.9 #34852

Closed

Merge remote-tracking branch 'upstream/master' into no-np-bool

7ed2e41

TomAugspurger merged commit 72aed3e into pandas-dev:master Jun 17, 2020

MatteoFelici pushed a commit to MatteoFelici/pandas that referenced this pull request Jun 18, 2020

Merge branch 'master' into format - fix CI fail pandas-dev#34835

38a6942

alfaro96 mentioned this pull request Jun 25, 2020

MNT More replacements of numpy aliases with built-in types scikit-learn/scikit-learn#17707

Merged

WillAyd deleted the no-np-bool branch June 30, 2020 16:05

simonjayhawkins mentioned this pull request Jan 7, 2021

DOC: np.bool -> np.bool_ #39019

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Replaced np.bool refs; fix CI failure #34835

Replaced np.bool refs; fix CI failure #34835

WillAyd commented Jun 16, 2020 •

edited by TomAugspurger

Loading

jbrockmendel commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

TomAugspurger Jun 17, 2020

TomAugspurger Jun 17, 2020

jorisvandenbossche Jun 17, 2020

jorisvandenbossche Jun 17, 2020

jreback commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020 •

edited

Loading

TomAugspurger commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

WillAyd commented Jun 17, 2020

WillAyd commented Jun 17, 2020

eric-wieser commented Jun 30, 2020

Replaced np.bool refs; fix CI failure #34835

Replaced np.bool refs; fix CI failure #34835

Conversation

WillAyd commented Jun 16, 2020 • edited by TomAugspurger Loading

jbrockmendel commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

TomAugspurger Jun 17, 2020

Choose a reason for hiding this comment

TomAugspurger Jun 17, 2020

Choose a reason for hiding this comment

jorisvandenbossche Jun 17, 2020

Choose a reason for hiding this comment

jorisvandenbossche Jun 17, 2020

Choose a reason for hiding this comment

jreback commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020 • edited Loading

TomAugspurger commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

TomAugspurger commented Jun 17, 2020

WillAyd commented Jun 17, 2020

WillAyd commented Jun 17, 2020

eric-wieser commented Jun 30, 2020

WillAyd commented Jun 16, 2020 •

edited by TomAugspurger

Loading

TomAugspurger commented Jun 17, 2020 •

edited

Loading