DEPR update dedeprecation message for np.ptp #28665 #6581 #30401

hasnain2808 · 2019-12-22T12:11:30Z

ref : Odd error message, concerning numpy using an alias versus numpy package name #28665 (comment) DEPR: Clean up list of deprecations from prior versions #6581
passes black pandas
passes git diff upstream/master -u -- "*.py" | flake8 --diff

…dev#6581

jbrockmendel · 2019-12-22T16:16:03Z

The fix here is not to update the warning but to remove Series.ptp entirely

hasnain2808 · 2019-12-22T17:00:58Z

Here in #28665
#28665 (comment)
It was suggested that till the time this resolved we should update the message

gfyoung · 2019-12-22T20:31:57Z

pandas/core/generic.py

@@ -10198,7 +10198,10 @@ def nanptp(values, axis=0, skipna=True):
            nmin = nanops.nanmin(values, axis, skipna)
            warnings.warn(
                "Method .ptp is deprecated and will be removed "
-                "in a future version. Use numpy.ptp instead.",
+                "in a future version. Use numpy.ptp instead."
+                "if you are already using numpy.ptp and still getting this message"


Suggested change

"if you are already using numpy.ptp and still getting this message"

"if you are already using numpy.ptp and still getting this message, "

gfyoung · 2019-12-22T20:32:45Z

pandas/core/generic.py

-                "in a future version. Use numpy.ptp instead.",
+                "in a future version. Use numpy.ptp instead."
+                "if you are already using numpy.ptp and still getting this message"
+                "please call to_numpy() to avoid this message in future calls"


Suggested change

"please call to_numpy() to avoid this message in future calls"

"please call to_numpy() to avoid this message in future calls. "

gfyoung · 2019-12-22T20:32:56Z

pandas/core/generic.py

+                "in a future version. Use numpy.ptp instead."
+                "if you are already using numpy.ptp and still getting this message"
+                "please call to_numpy() to avoid this message in future calls"
+                "eg : np.ptp(pd.Series([1, 2, 3]).to_numpy())",


Suggested change

"eg : np.ptp(pd.Series([1, 2, 3]).to_numpy())",

"For example: np.ptp(pd.Series([1, 2, 3]).to_numpy())",

jreback

agree with @jbrockmendel let's just remove this entirely. it has been deprecated for a while and we are removing code for 1.0

…dev#6581 requested changes made

hasnain2808 · 2019-12-23T16:58:46Z

agree with @jbrockmendel let's just remove this entirely. it has been deprecated for a while and we are removing code for 1.0

so, for now, I have updated the message,
I believe it will take some time for getting rid of np,ptp() till then this message should help.

jreback · 2019-12-23T17:32:58Z

agree with @jbrockmendel let's just remove this entirely. it has been deprecated for a while and we are removing code for 1.0

so, for now, I have updated the message,
I believe it will take some time for getting rid of np,ptp() till then this message should help.

no, for 1.0 we are going to remove this code, if you want to repurpose this PR to do it great.

hasnain2808 · 2019-12-23T17:37:51Z

Ohhk no isssues
I would love to do it
As I am new to the codebase it would be really nice if you can give me some pointers around achieving this

jreback · 2019-12-23T17:54:34Z

Ohhk no isssues
I would love to do it
As I am new to the codebase it would be really nice if you can give me some pointers around achieving this

ok, you remove the code & associated tests, see https://dev.pandas.io/docs/whatsnew/v1.0.0.html#removal-of-prior-version-deprecations-changes for things that we have removed, so add an entry liek this

hasnain2808 · 2019-12-23T21:09:30Z

so I removed code and tests for ptp
remove the part where series only functions are added
removed the entire series only function because ptp was the only series only function
deleted tests for pandas ptp function
but the test written for np.ptp gives me a future warning

`
____________________________________________________________ TestSeriesAnalytics.test_ptp ____________________________________________________________

self = <pandas.tests.series.test_analytics.TestSeriesAnalytics object at 0x7f666e0267b8>

def test_ptp(self):
    # GH21614
    N = 1000
    arr = np.random.randn(N)
    ser = Series(arr)
    with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):

      assert np.ptp(ser) == np.ptp(arr)

pandas/tests/series/test_analytics.py:861:

self = <contextlib._GeneratorContextManager object at 0x7f666e03ad30>, type = None, value = None, traceback = None

def __exit__(self, type, value, traceback):
    if type is None:
        try:

          next(self.gen)

E AssertionError: Did not see expected warning of class 'FutureWarning'

/usr/lib/python3.6/contextlib.py:88: AssertionError
================================================ 1 failed, 163 passed, 1 skipped, 3 xfailed in 1.96s =================================================
`

jbrockmendel · 2019-12-23T21:13:31Z

pandas/core/generic.py

-            ``numpy.ndarray`` method ``ptp``.""",
-            nanptp,
-        )
+    # @classmethod


delete all of this

jbrockmendel · 2019-12-23T21:13:45Z

pandas/core/series.py

@@ -4390,7 +4390,7 @@ def to_period(self, freq=None, copy=True):
    ["index"], docs={"index": "The index (axis labels) of the Series."},
 )
 Series._add_numeric_operations()
-Series._add_series_only_operations()
+# Series._add_series_only_operations()


delete. basically never leave commented-out code in a PR

actually made this commit to get some help on getting this future warning

so I removed code and tests for ptp
remove the part where series only functions are added
removed the entire series only function because ptp was the only series only function
deleted tests for pandas ptp function
but the test written for np.ptp gives me a future warning

`
____________________________________________________________ TestSeriesAnalytics.test_ptp ____________________________________________________________

self = <pandas.tests.series.test_analytics.TestSeriesAnalytics object at 0x7f666e0267b8>

def test_ptp(self): # GH21614 N = 1000 arr = np.random.randn(N) ser = Series(arr) with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):

assert np.ptp(ser) == np.ptp(arr)

pandas/tests/series/test_analytics.py:861:

self = <contextlib._GeneratorContextManager object at 0x7f666e03ad30>, type = None, value = None, traceback = None

def __exit__(self, type, value, traceback): if type is None: try:

next(self.gen)

E AssertionError: Did not see expected warning of class 'FutureWarning'

/usr/lib/python3.6/contextlib.py:88: AssertionError
================================================ 1 failed, 163 passed, 1 skipped, 3 xfailed in 1.96s =================================================
`

Remove the assert_produces_warning, since it shouldn't anymore.

it works but can someone please explain to me the reason behind it so that I learn

pandas/tests/series/test_analytics.py

hasnain2808 · 2019-12-24T05:13:35Z

tests are not failing at local no clue why azure pipelines are failing

`
(pandas-dev) moha@moha-HP-Pavilion-Notebook:~/projects/pandas-hasnain2808$ pytest pandas/tests/series/test_analytics.py
============================= test session starts ==============================
platform linux -- Python 3.6.9, pytest-5.3.2, py-1.8.0, pluggy-0.13.1
rootdir: /home/moha/projects/pandas-hasnain2808, inifile: setup.cfg
plugins: cov-2.8.1, forked-1.1.3, xdist-1.30.0, hypothesis-4.54.0
collected 168 items

pandas/tests/series/test_analytics.py ..........................s....... [ 20%]
........................................................................ [ 63%]
..............................................x.....x.....x... [100%]

================== 164 passed, 1 skipped, 3 xfailed in 3.26s ===================
`

This reverts commit 5d8d35f.

This reverts commit 86c11c9.

This reverts commit b28bc3a.

This reverts commit add3e23.

This reverts commit 3df0283.

This reverts commit 08a5f3e.

This reverts commit d9e955e.

…andas-dev#6581 requested changes made" This reverts commit d89d7f6.

…nd why azure build failing

jbrockmendel · 2019-12-24T21:24:01Z

What was happening in the failing builds is that when we call np.ptp(ser) numpy first checks if ser has a ptp method, in which case it calls that. This PR removes (removed) that method, so numpy is falling back to whatever it tries next, probably involving either __array__ or __array_func__ or something. Whatever that something is, it's returning a Series when it should return a scalar.

hasnain2808 · 2019-12-25T01:34:19Z

What was happening in the failing builds is that when we call np.ptp(ser) numpy first checks if ser has a ptp method, in which case it calls that. This PR removes (removed) that method, so numpy is falling back to whatever it tries next, probably involving either __array__ or __array_func__ or something. Whatever that something is, it's returning a Series when it should return a scalar.

So I actually had tried what is numpy returning on my local machine. Over there it's returning float
And also it would be nice if you could guide me a bit more towards the solution of this numpy issue

…e and values of ptp to understand why azure build failing

hasnain2808 · 2019-12-25T05:46:03Z

What was happening in the failing builds is that when we call np.ptp(ser) numpy first checks if ser has a ptp method, in which case it calls that. This PR removes (removed) that method, so numpy is falling back to whatever it tries next, probably involving either __array__ or __array_func__ or something. Whatever that something is, it's returning a Series when it should return a scalar.

As I am unable to reproduce this on my local machine I tried to push a commit with the print statements which are ignored by the pipeline.
Also only builds on some platforms are failing not all
link to the azure build

mac os python 3.6
linux python 3.6 minimum version
linux python 3.6 32 bit
windows python 3.7 np 141

checking the logs show me that the series returned by numpy is all True boolean series

2019-12-25T01:57:55.8017700Z self = 0 True
2019-12-25T01:57:55.8017950Z 1 True
2019-12-25T01:57:55.8018210Z 2 True
2019-12-25T01:57:55.8018480Z 3 True
2019-12-25T01:57:55.8018720Z 4 True
2019-12-25T01:57:55.8018980Z ...
2019-12-25T01:57:55.8019230Z 995 True
2019-12-25T01:57:55.8019480Z 996 True
2019-12-25T01:57:55.8019740Z 997 True
2019-12-25T01:57:55.8019990Z 998 True
2019-12-25T01:57:55.8020250Z 999 True
2019-12-25T01:57:55.8020500Z Length: 1000, dtype: bool
2019-12-25T01:57:55.8020730Z

if it would have been returning a series with just the correct values it would have been simple to extract it, but the returned series is not correct.

hasnain2808 · 2019-12-25T06:39:38Z

At last all tests passing
Requesting some reviewer to check it

jreback · 2019-12-25T23:52:07Z

thanks for the effort @hasnain2808 I had already merged #30458

happy to take a contribution on other issues.

update dedeprecation message for np.ptp DEPR pandas-dev#28665 pandas-…

da3c79a

…dev#6581

hasnain2808 changed the title ~~update dedeprecation message for np.ptp DEPR #28665 #6581~~ DEPR update dedeprecation message for np.ptp #28665 #6581 Dec 22, 2019

gfyoung added the Deprecate Functionality to remove in pandas label Dec 22, 2019

gfyoung reviewed Dec 22, 2019

View reviewed changes

jreback requested changes Dec 23, 2019

View reviewed changes

update dedeprecation message for np.ptp DEPR pandas-dev#28665 pandas-…

d89d7f6

…dev#6581 requested changes made

remove code for ptp()

d9e955e

jbrockmendel reviewed Dec 23, 2019

View reviewed changes

pandas/tests/series/test_analytics.py Outdated Show resolved Hide resolved

hasnain2808 added 3 commits December 24, 2019 02:51

remove comments)

08a5f3e

remove comments

3df0283

solve errors

add3e23

hasnain2808 closed this Dec 24, 2019

hasnain2808 reopened this Dec 24, 2019

hasnain2808 added 3 commits December 24, 2019 12:18

test at which command azure pipelines failing

b28bc3a

use numpy.array_equal instead of ==

86c11c9

use numpy.array_equal instead of ==

5d8d35f

hasnain2808 closed this Dec 24, 2019

hasnain2808 reopened this Dec 24, 2019

hasnain2808 added 9 commits December 25, 2019 01:32

Revert "use numpy.array_equal instead of =="

c9ee01a

This reverts commit 5d8d35f.

Revert "use numpy.array_equal instead of =="

0610de5

This reverts commit 86c11c9.

Revert "test at which command azure pipelines failing"

f1e82ab

This reverts commit b28bc3a.

Revert "solve errors"

d8b649e

This reverts commit add3e23.

Revert "remove comments"

eeba706

This reverts commit 3df0283.

Revert "remove comments)"

0c19cd3

This reverts commit 08a5f3e.

Revert "remove code for ptp()"

ab8343a

This reverts commit d9e955e.

Revert "update dedeprecation message for np.ptp DEPR pandas-dev#28665 p…

69d1c0a

…andas-dev#6581 requested changes made" This reverts commit d89d7f6.

removed ptp code and tests and also put a print statement to understa…

8c0bb01

…nd why azure build failing

removed ptp code and tests and also put print statements to print typ…

cc66afe

…e and values of ptp to understand why azure build failing

casting pandas to numpy before calling np.ptp()

d9e9ac7

hasnain2808 mentioned this pull request Dec 25, 2019

DEPR: remove ptp #30458

Merged

jreback added this to the 1.0 milestone Dec 25, 2019

jreback closed this Dec 25, 2019

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

DEPR update dedeprecation message for np.ptp #28665 #6581 #30401

DEPR update dedeprecation message for np.ptp #28665 #6581 #30401

hasnain2808 commented Dec 22, 2019 •

edited by gfyoung

jbrockmendel commented Dec 22, 2019

hasnain2808 commented Dec 22, 2019 •

edited

gfyoung Dec 22, 2019

hasnain2808 Dec 23, 2019

gfyoung Dec 22, 2019

hasnain2808 Dec 23, 2019

gfyoung Dec 22, 2019

hasnain2808 Dec 23, 2019

jreback left a comment

hasnain2808 commented Dec 23, 2019

jreback commented Dec 23, 2019

hasnain2808 commented Dec 23, 2019

jreback commented Dec 23, 2019

hasnain2808 commented Dec 23, 2019 •

edited

jbrockmendel Dec 23, 2019

hasnain2808 Dec 23, 2019

jbrockmendel Dec 23, 2019

hasnain2808 Dec 23, 2019 •

edited

hasnain2808 Dec 23, 2019 •

edited

jbrockmendel Dec 23, 2019

hasnain2808 Dec 24, 2019

hasnain2808 commented Dec 24, 2019

jbrockmendel commented Dec 24, 2019

hasnain2808 commented Dec 25, 2019

hasnain2808 commented Dec 25, 2019 •

edited

hasnain2808 commented Dec 25, 2019

jreback commented Dec 25, 2019

	"if you are already using numpy.ptp and still getting this message"
	"if you are already using numpy.ptp and still getting this message, "

	"please call to_numpy() to avoid this message in future calls"
	"please call to_numpy() to avoid this message in future calls. "

	"eg : np.ptp(pd.Series([1, 2, 3]).to_numpy())",
	"For example: np.ptp(pd.Series([1, 2, 3]).to_numpy())",

DEPR update dedeprecation message for np.ptp #28665 #6581 #30401

DEPR update dedeprecation message for np.ptp #28665 #6581 #30401

Conversation

hasnain2808 commented Dec 22, 2019 • edited by gfyoung

jbrockmendel commented Dec 22, 2019

hasnain2808 commented Dec 22, 2019 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jreback left a comment

Choose a reason for hiding this comment

hasnain2808 commented Dec 23, 2019

jreback commented Dec 23, 2019

hasnain2808 commented Dec 23, 2019

jreback commented Dec 23, 2019

hasnain2808 commented Dec 23, 2019 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

hasnain2808 Dec 23, 2019 • edited

Choose a reason for hiding this comment

hasnain2808 Dec 23, 2019 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

hasnain2808 commented Dec 24, 2019

jbrockmendel commented Dec 24, 2019

hasnain2808 commented Dec 25, 2019

hasnain2808 commented Dec 25, 2019 • edited

hasnain2808 commented Dec 25, 2019

jreback commented Dec 25, 2019

hasnain2808 commented Dec 22, 2019 •

edited by gfyoung

hasnain2808 commented Dec 22, 2019 •

edited

hasnain2808 commented Dec 23, 2019 •

edited

hasnain2808 Dec 23, 2019 •

edited

hasnain2808 Dec 23, 2019 •

edited

hasnain2808 commented Dec 25, 2019 •

edited