MAINT: Sync with randomgen changes by bashtage · Pull Request #2 · mattip/numpy

bashtage · 2019-04-01T15:34:43Z

Synchronize with changes in randomgen

This is a very large PR that brings in all of the changes from randomgen.

The most important are:

Remove LegacyGenerator
Add a stand-alone RandomState that will be frozen, following @rkern's suggestion.
Quite a few small bug fixes -- coverage is near 100% now.
SOme refactoring to remove duplicated code, expecially in the basic RNGs.

I'm not sure this is the best way tot get these in. I fyou think it is better to pull from randomgen then please pull these in the best way.

Synchronize with changes in randomgen

mattip · 2019-04-01T15:52:29Z

Absolutely better than me trying to merge it from bashtage/randomgen.

bashtage · 2019-04-01T16:38:10Z

It is failing for some strange reason that seems unrelated to these changes:

https://travis-ci.org/bashtage/numpy/builds/514226868

=================================== FAILURES ===================================
______________________ TestAlmostEqual.test_error_message ______________________
self = <numpy.testing.tests.test_utils.TestAlmostEqual object at 0x7fa7b9975c18>
    def test_error_message(self):
        """Check the message is formatted correctly for the decimal value.
           Also check the message when input includes inf or nan (gh12200)"""
        x = np.array([1.00000000001, 2.00000000002, 3.00003])
        y = np.array([1.00000000002, 2.00000000003, 3.00004])
    
        # Test with a different amount of decimal digits
        with pytest.raises(AssertionError) as exc_info:
            self._assert_func(x, y, decimal=12)
        msgs = str(exc_info.value).split('\n')
        assert_equal(msgs[3], 'Mismatch: 100%')
>       assert_equal(msgs[4], 'Max absolute difference: 1.e-05')
E       AssertionError: 
E       Items are not equal:
E        ACTUAL: 'Max absolute difference: 9.999999999621423e-06'
E        DESIRED: 'Max absolute difference: 1.e-05'

mattip · 2019-04-01T16:59:58Z

Strange. Does that reproduce locally?

bashtage · 2019-04-01T17:01:41Z

No, passes locally.

bashtage · 2019-04-01T17:33:12Z

It does repro locally, on both Windows and Linux. I was only running tests under random, and so I didn't see it.

bashtage · 2019-04-01T17:46:13Z

It looks like the precision is being changed somewhere to that it isn't rounding right. This is a formatted string.
`ACTUAL: 'Max absolute difference: 9.999999999621423e-06'

I also tried rebasing which didn't fix things.

bashtage · 2019-04-01T17:46:40Z

It should be 1.00000e-05

mattip · 2019-04-01T17:48:02Z

Interesting, it is similar to an issue we have on linux32 wheel builds where something is reducing floating point accuracy. A compiler glitch? Some assemble code somewhere changing registers?

Incorporate testing of edge cases into main tests Rename test files to describe their purpose

bashtage · 2019-04-01T17:50:05Z

It is my fault - I left a np.set_printiontions(precision=20) in a test.

mattip · 2019-04-01T17:56:25Z

@@ -6,6 +7,7 @@ from randomgen.common cimport *
 from randomgen.distributions cimport random_gauss_zig
 from randomgen.xoroshiro128 import Xoroshiro128


from numpy.random.radomgen

mattip · 2019-04-01T17:59:28Z

        random_integers(low, high=None, size=None)

-        Random integers of type np.int64 between `low` and `high`, inclusive.
+        Random integers of type np.int between `low` and `high`, inclusive.


On 32 bit systems this gives int32?

Yes, also on all bitness Windows. This function is deprecated and IMO should be removed before this transition.

Import import locatiosn to reflect numpy paths

bashtage · 2019-04-01T20:03:13Z

This is final from my end. I think it gets it very close, although I know there are still docs and I'm sure other things.

Coverage is pretty good -- should be the same as randomgen:

https://codecov.io/gh/bashtage/randomgen

It is in the low 90s but coverage flags all def lines as not run even when they are, so I suspect that it is probably 98% or so, with only some hard to reach defensive code not covered.

mattip · 2019-04-01T20:25:16Z

Thanks. Merging like this is much simpler than hand-importing from randomgen

commit 9c833bed5879d77e625556260690c349de18b433 Author: Thomas Li <[email protected]> Date: Wed Nov 17 16:21:27 2021 -0800 Add Windows config to GHA update script [wheel build] typo [wheel build] fix typo? [wheel build] fix linux builds? [wheel build] typo [wheel build] add license and pin to windows 2016 skip tests [wheel build] pin to windows 2019 instead [wheel build] try to find out the error on windows [wheel build] maybe fix? [wheel build] maybe fix? [wheel build] fix? [wheel build] cleanup [wheel build] Add Windows config to GHA update script [wheel build] typo [wheel build] fix typo? [wheel build] fix linux builds? [wheel build] typo [wheel build] add license and pin to windows 2016 skip tests [wheel build] pin to windows 2019 instead [wheel build] try to find out the error on windows [wheel build] maybe fix? [wheel build] maybe fix? [wheel build] fix? [wheel build] cleanup [wheel build] Update LICENSE_win32.txt Update LICENSE_win32.txt Add Windows config to GHA update script [wheel build] typo [wheel build] fix typo? [wheel build] fix linux builds? [wheel build] typo [wheel build] add license and pin to windows 2016 skip tests [wheel build] pin to windows 2019 instead [wheel build] try to find out the error on windows [wheel build] maybe fix? [wheel build] maybe fix? [wheel build] fix? [wheel build] cleanup [wheel build] Update LICENSE_win32.txt Update LICENSE_win32.txt Update cibw_test_command.sh commit 4bd12df Author: Thomas Li <[email protected]> Date: Mon Nov 15 17:28:47 2021 -0800 # This is a combination of 14 commits. # This is the 1st commit message: Add Windows config to GHA # This is the commit message #2: update script [wheel build] # This is the commit message #3: typo [wheel build] # This is the commit message #4: fix typo? [wheel build] # This is the commit message #5: fix linux builds? [wheel build] # This is the commit message #6: typo [wheel build] # This is the commit message #7: add license and pin to windows 2016 # This is the commit message #8: skip tests [wheel build] # This is the commit message #9: pin to windows 2019 instead [wheel build] # This is the commit message #10: try to find out the error on windows [wheel build] # This is the commit message #11: maybe fix? [wheel build] # This is the commit message #12: maybe fix? [wheel build] # This is the commit message #13: fix? [wheel build] # This is the commit message #14: cleanup [wheel build]

MAINT: Sync with randomgen changes

a884450

Synchronize with changes in randomgen

TST: Incorporate edge tests

ce5857d

Incorporate testing of edge cases into main tests Rename test files to describe their purpose

bashtage force-pushed the randomgen-update branch from 9e87187 to ce5857d Compare April 1, 2019 17:49

mattip reviewed Apr 1, 2019

View reviewed changes

Comment thread numpy/random/randomgen/examples/cython/extending.pyx Outdated

mattip reviewed Apr 1, 2019

View reviewed changes

MAINT/DOC: Update import locations

ac66298

Import import locatiosn to reflect numpy paths

mattip merged commit ee3593d into mattip:randomgen Apr 1, 2019

bashtage deleted the randomgen-update branch April 15, 2019 15:49

mattip pushed a commit that referenced this pull request Aug 23, 2019

Added email, fixed typos (#2)

ea31b08

mattip pushed a commit that referenced this pull request Oct 7, 2020

TST: Attempt #2 at fixing the int-related failure on 32-bit systems

e67bbd2

mattip pushed a commit that referenced this pull request Apr 16, 2021

change to suitable raise from and add line breaks #2

fdeb979

		@@ -6,6 +7,7 @@ from randomgen.common cimport *
		from randomgen.distributions cimport random_gauss_zig
		from randomgen.xoroshiro128 import Xoroshiro128

Conversation

bashtage commented Apr 1, 2019

Uh oh!

mattip commented Apr 1, 2019

Uh oh!

bashtage commented Apr 1, 2019

Uh oh!

mattip commented Apr 1, 2019

Uh oh!

bashtage commented Apr 1, 2019

Uh oh!

bashtage commented Apr 1, 2019

Uh oh!

bashtage commented Apr 1, 2019

Uh oh!

bashtage commented Apr 1, 2019

Uh oh!

mattip commented Apr 1, 2019

Uh oh!

bashtage commented Apr 1, 2019

Uh oh!

Uh oh!

mattip Apr 1, 2019

Choose a reason for hiding this comment

Uh oh!

mattip Apr 1, 2019

Choose a reason for hiding this comment

Uh oh!

bashtage Apr 1, 2019

Choose a reason for hiding this comment

Uh oh!

bashtage commented Apr 1, 2019

Uh oh!

mattip commented Apr 1, 2019

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants