MAINT: constants: recalculation of exact constants #17584

jakobjakobson13 · 2022-12-11T19:49:33Z

Reference issue

Closes gh-11341, closes gh-11345 and closes gh-14467.

What does this implement/fix?

Exact values of the CODATA constant set are calculated for bigger precision

Additional information

This is just a follow up on the abandoned gh-11345 and credits for the working parts go to @pv (The broken bits are my mistake.)

see scipy#11345 as reference

scipy/constants/tests/test_codata.py

tupui

Thanks for the work! It's in good shape now, just some minor things with the linter.

scipy/constants/tests/test_codata.py

jakobjakobson13 · 2022-12-13T17:03:36Z

Thanks for the work! It's in good shape now, just some minor things with the linter.

@tupui Could you take a look at the linter/formatting issues? I don´t know how to handle them.

tupui

I did a few suggestions to fix the linter. For some lines too long in _codata it's ok to let it complain.

scipy/constants/tests/test_codata.py

scipy/constants/_codata.py

scipy/constants/tests/test_codata.py

tupui · 2022-12-13T17:48:26Z

scipy/constants/_codata.py

+        elif is_exact:
+            exact[name] = val
+        else:
+            assert not is_truncated


Here and elsewhere:
Is this for control flow? If yes, could you replace these with proper exceptions? The issue is that assertions are missing if someone use python -O

Co-authored-by: Pamphile Roy <roy.pamphile@gmail.com>

tupui · 2022-12-23T18:14:00Z

scipy/constants/_codata.py

+                raise Exception("Parsing error")
+        except Warning as w:
+            raise Warning(f"{name} is truncated but not exact.") from w


About the exceptions, we should try to raise informative types of exceptions for the users, instead of Warning, it's more something like a ValueError that we might need.

For internal exceptions, we should also be more specific and not use a bare Exception. This to help prevent catching something else.

Lastly and this is a bit more work, we should test any new exception which we add. Our current policy is to do the following in Pytest:

message = r"Message in the exception" with pytest.raises(ValueError, match=message): ... # code that will raise

Okay, I'll try to get to it at some other day.

Thank you 🙇 No rush at all.

jakobjakobson13 · 2023-01-03T20:29:37Z

The tests show the intended error messages but I admit that the whole pull request became such a mess that I'm not sure if I made it better or worse. So, before you merge it, please perform a very, very thorough review.

jakobjakobson13 · 2023-11-19T14:00:42Z

@lucascolley
Regarding your latest post in #11345: Unfortunately I won´t have the time to check the pull requests thoroughly in the next weeks.

lucascolley · 2023-11-19T14:51:04Z

Unfortunately I won´t have the time to check the pull requests thoroughly in the next weeks.

Thanks a lot for the quick response. I've bumped the milestone on the assumption that no maintainers are looking to give the "very, very thorough review" and finish this in the next few weeks.

tylerjereddy · 2024-05-26T23:00:40Z

no maintainers are looking to give the "very, very thorough review" and finish this in the next few weeks.

Probably not, and I'm hoping we don't have to wait weeks before I branch 1.14.0, so I'm going to bump the milestone again, apologies!

jakobjakobson13 · 2024-05-27T14:39:10Z

Probably not, and I'm hoping we don't have to wait weeks before I branch 1.14.0, so I'm going to bump the milestone again, apologies!

Don't worry, I'm currently occupied with other things with a higher priority for me.

mdhaber · 2024-09-30T01:16:02Z

scipy/constants/tests/test_codata.py

+from numpy.testing import (
+    assert_equal,
+    assert_,
+    assert_almost_equal,
+    assert_raises,
+)
+from scipy.constants import _codata


Let's not make unnnecessary formatting changes.

mdhaber · 2024-09-30T01:17:47Z

scipy/constants/_codata.py

+            elif is_exact:
+                exact[name] = val
+            elif is_truncated:
+                raise Warning("Parsing error")


Why raisea Warning? Should this warn or raise an exception?

mdhaber · 2024-09-30T01:18:08Z

scipy/constants/_codata.py


-physical_constants: dict[str, tuple[float, str, float]] = {}
+exact2022 = exact2018


I haven't checked whether this is true, but for now...

mdhaber · 2024-09-30T01:19:01Z

scipy/constants/_codata.py

+    h = exact['Planck constant']
+    e = exact['elementary charge']
+    k = exact['Boltzmann constant']
+    N_A = exact['Avogadro constant']


The original PR defined some useful constants to reduce duplication. I'll add them back.

mdhaber · 2024-09-30T01:21:57Z

scipy/constants/_codata.py

+    alpha_W = (3 + lambertw(-3 * math.exp(-3))).real
+    x_W = (5 + lambertw(-5 * math.exp(-5))).real


The original PR defined numerical values. Fortunately, lambertw gives exactly the same result as mpmath, but it would probably be safer to pre-calculate these.

[lint only]

mdhaber · 2024-09-30T02:22:13Z

Thanks for working on this @jakobjakobson13.

After merging main into gh-11345 and this PR, I checked this against the original, gh-11345. I've decided to close this one in favor of gh-11345 for a few reasons:

It appears that this was a manual recreation and adjustment of MAINT: constants: revise way 'exact' values are recomputed #11345. The PR that gets merged should include the original commits.
The raise Warnings do not seem to be an improvement over the assertions in the original PR. Neither are ideal, but the intent of the original assertions was more clear to me.
There were unrelated formatting changes here; these make the diff harder to review.
There were some good ideas from the original PR that were left out of this one (e.g. definition of hbar and reuse)

There were also some good ideas from this one that I merged into gh-11345 in 4a91f45, so when that gets merged, we'll add you as a coauthor.

jakobjakobson13 added 3 commits December 11, 2022 20:39

Recalculate exact values

24f69ba

see scipy#11345 as reference

Adapt test to recalculated values

3e014de

fix intendation

d03f8cf

jakobjakobson13 changed the title ~~MAINT: Recalculation of exact values~~ MAINT: Recalculation of exact constants Dec 11, 2022

Add import of lambertw

ca58e29

tupui added enhancement A new feature or improvement scipy.constants labels Dec 11, 2022

tupui added this to the 1.11.0 milestone Dec 11, 2022

jakobjakobson13 added 5 commits December 11, 2022 21:45

Add import of exp()

4c36a03

Replace hbar by h/(2*pi)

688a2a5

Replace G_0 by (2 * e**2 / h)

29df3fe

Replace R by (N_A * k)

46475d7

Fix parsing functions

27d1ece

tupui reviewed Dec 12, 2022

View reviewed changes

scipy/constants/tests/test_codata.py Outdated Show resolved Hide resolved

jakobjakobson13 added 6 commits December 12, 2022 18:24

Fix tests

46e3630

Fix tests

0925534

Fix tests

857561f

Fix linter errors

5e67c49

Use only real part for Wien law calculation

5e2d6c0

Delete wrong string

7e7320f

jakobjakobson13 marked this pull request as ready for review December 12, 2022 22:19

tupui reviewed Dec 13, 2022

View reviewed changes

scipy/constants/tests/test_codata.py Show resolved Hide resolved

tupui reviewed Dec 13, 2022

View reviewed changes

jakobjakobson13 and others added 6 commits December 13, 2022 18:55

Fix linter errors

668fa86

Co-authored-by: Pamphile Roy <roy.pamphile@gmail.com>

Fix linter error

10d4e9b

Co-authored-by: Pamphile Roy <roy.pamphile@gmail.com>

Fix formatting

b2a8756

Co-authored-by: Pamphile Roy <roy.pamphile@gmail.com>

Fix formatting

7e0aa22

Co-authored-by: Pamphile Roy <roy.pamphile@gmail.com>

Fix formatting

491ba8a

Co-authored-by: Pamphile Roy <roy.pamphile@gmail.com>

Fix formatting

272fd93

Co-authored-by: Pamphile Roy <roy.pamphile@gmail.com>

tupui reviewed Dec 23, 2022

View reviewed changes

jakobjakobson13 added 2 commits December 23, 2022 18:22

Delete blank lines

88a6df4

Add tests

03cc728

jakobjakobson13 added 2 commits January 3, 2023 22:44

Add missing imports

11f2253

Change exception to warning

fc49eb1

h-vetinari modified the milestones: 1.11.0, 1.12.0 May 20, 2023

lucascolley modified the milestones: 1.12.0, 1.13.0 Nov 19, 2023

lucascolley modified the milestones: 1.13.0, 1.14.0 Mar 12, 2024

jakobjakobson13 mentioned this pull request Apr 9, 2024

MAINT: constants: slim down module, part 2 #20430

Closed

lucascolley changed the title ~~MAINT: Recalculation of exact constants~~ MAINT: constants: recalculation of exact constants May 18, 2024

tylerjereddy modified the milestones: 1.14.0, 1.15.0 May 26, 2024

mdhaber added 4 commits September 29, 2024 16:29

Merge remote-tracking branch 'upstream/main' into new_exact_constants

3a0071e

MAINT: constants: fix merge mistakes

640d766

STY: constants: double to single quotes

99a3fc8

STY: constants: revert unnecessary formatting changes

278b39a

mdhaber reviewed Sep 30, 2024

View reviewed changes

TST: constants: add back new tests

173ffcf

[lint only]

mdhaber added a commit to pv/scipy-work that referenced this pull request Sep 30, 2024

MAINT: constants: update with ideas from scipygh-17584

4a91f45

mdhaber closed this Sep 30, 2024

mdhaber mentioned this pull request Sep 30, 2024

MAINT: constants: revise way 'exact' values are recomputed #11345

Merged


		physical_constants: dict[str, tuple[float, str, float]] = {}
		exact2022 = exact2018

		alpha_W = (3 + lambertw(-3 * math.exp(-3))).real
		x_W = (5 + lambertw(-5 * math.exp(-5))).real

Uh oh!

MAINT: constants: recalculation of exact constants #17584

MAINT: constants: recalculation of exact constants #17584

Uh oh!

Conversation

jakobjakobson13 commented Dec 11, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Reference issue

What does this implement/fix?

Additional information

Uh oh!

Uh oh!

tupui left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

jakobjakobson13 commented Dec 13, 2022

Uh oh!

tupui left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

tupui Dec 13, 2022

Choose a reason for hiding this comment

Uh oh!

tupui Dec 23, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jakobjakobson13 Dec 23, 2022

Choose a reason for hiding this comment

Uh oh!

tupui Dec 23, 2022

Choose a reason for hiding this comment

Uh oh!

jakobjakobson13 commented Jan 3, 2023

Uh oh!

jakobjakobson13 commented Nov 19, 2023

Uh oh!

lucascolley commented Nov 19, 2023

Uh oh!

tylerjereddy commented May 26, 2024

Uh oh!

jakobjakobson13 commented May 27, 2024

Uh oh!

mdhaber Sep 30, 2024

Choose a reason for hiding this comment

Uh oh!

mdhaber Sep 30, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

mdhaber Sep 30, 2024

Choose a reason for hiding this comment

Uh oh!

mdhaber Sep 30, 2024

Choose a reason for hiding this comment

Uh oh!

mdhaber Sep 30, 2024

Choose a reason for hiding this comment

Uh oh!

mdhaber commented Sep 30, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

jakobjakobson13 commented Dec 11, 2022 •

edited

Loading

tupui Dec 23, 2022 •

edited

Loading

mdhaber Sep 30, 2024 •

edited

Loading

mdhaber commented Sep 30, 2024 •

edited

Loading