split out a separate Python package primecountpy

[dimpase]

This is a followup to #25009, and also resolves the bugs mentioned in #24960.

Here is my attempt (WIP) to create such a package:
https://github.com/dimpase/primecountpy

On PyPI: https://pypi.org/project/primecountpy/

Depends on #25009

CC: @mkoeppe

Component: packages: standard

Author: Dima Pasechnik

Branch/Commit: 74b3845

Reviewer: Matthias Koeppe

Issue created by migration from https://trac.sagemath.org/ticket/32894

[mkoeppe]

comment:2

As you know, https://pypi.org/project/primecount/ is already taken...

[dimpase]

comment:3

I must say I don't quite know what exactly pypi package name is, and whether anything beyond changes in MANIFEST.in and setup.py is needed.
Can the module name still be the same, primecount ?

We'd call it pyprimecount instead.

[mkoeppe]

comment:4

I explained it recently in https://groups.google.com/g/sage-devel/c/p4MO6L95s7E/m/yZF93ewIBAAJ, take a look

[dimpase]

comment:5

OK, thanks. Another thing I don't understand is how the C++ library prinecount is meant to be provided. It's not something too standard to assume it's just installed.

Does it mean to be a non-python dependency of Sage? Or of (py)prinecount?

[mkoeppe]

comment:6

When pyprimecount is installed from a source distribution, it will just assume that the necessary library is available. Python packages do not have a way to declare non-Python dependencies.

Later one may consider to build wheels that include a copy of the shared library.
This is what for example numpy does regarding BLAS: The source distribution certainly does not contain a copy of BLAS, but the wheels ship with a copy of the OpenBLAS shared library.

[mkoeppe]

comment:7

On the Sage side, there will be two SPKGs: One for primecount, one for pyprimecount. This mirrors what is done, for example, for ppl and pplpy. The SPKG dependencies of pyprimecount include primecount; so when pyprimecount is pip-installed, the library has already been installed.

[dimpase]

comment:8

after a very careful consideration, decided to rename it primecountpy.

[dimpase]

comment:10

OK, so prinecountpy works (needs setting up proper tests, etc), but other than that, what else is needed on that side?Some pypi/pip magic?

(on Sage's side clearly code to be removed and replaced by hopefully clear something...)

[mkoeppe]

comment:11

when primecountpy is ready, use setup.py sdist to create an sdist and upload to PyPI with twine

Then on the Sage side, use ./sage -package create primecountpy --pypi

[mkoeppe]

Dependencies: #25009

[dimpase]

comment:13

OK, here it is: https://pypi.org/project/primecountpy/

[dimpase]

Last 10 new commits:

c655136	remove sagemath-primecount declaration
9ca1f67	misc config changes for primesieve
d8e2cda	adjust deps,remove stuff from cfg and toml
3f1e441	no need for CMAKE_INSTALL_PREFIX (it's in sdh_cmake)
abce309	flyby patch: remove CMAKE_INSTALL_PREFIX, as it's set in sdh_cmake
a6de7f4	check external primesieve with cmake
733cebc	prime_pi and legendre_phi calling primecount
4f77073	allow for slightly different Overflow messages in doctest
f7a0eed	spkg primecountpy
9e154ad	use primecountpy instead of built-in cython interface

[dimpase]

Branch: u/dimpase/packages/primecountpy

[dimpase]

Author: Dima Pasechnik

[dimpase]

Commit: 9e154ad

[dimpase]

comment:15

it works, apparently.

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

f3d12bc

remove primecount from various lists of optional spkgs

[sagetrac-git]

Changed commit from 9e154ad to f3d12bc

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. New commits:

d950913

fedora 30 doesn't have primecount pkg

[sagetrac-git]

Changed commit from f3d12bc to d950913

[mkoeppe]

comment:19

this should go to #25009

[isuruf]

comment:40

Can you add primesieve conda package too?

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1 and set ticket back to needs_review. New commits:

8ef586a

primesieve is on conda, too

[sagetrac-git]

Changed commit from 5495982 to 8ef586a

[dimpase]

comment:42

Replying to @isuruf:

Can you add primesieve conda package too?

done.

[vbraun]

comment:43

Test fail, e.g.

File "src/sage/plot/plot.py", line 1753, in sage.plot.plot.plot
Failed example:
    plot(lambda x: prime_pi(x), (x, 1, 100), exclude=jumps)
Expected:
    Graphics object consisting of 26 graphics primitives
Got:
    verbose 0 (3839: plot.py, generate_plot_points) WARNING: When plotting, failed to evaluate function at 248 points.
    verbose 0 (3839: plot.py, generate_plot_points) Last error message: 'unable to simplify to float approximation'
    Graphics object consisting of 0 graphics primitives
**********************************************************************
1 item had failures:
   1 of 147 in sage.plot.plot.plot
    [423 tests, 1 failure, 28.78 s]
----------------------------------------------------------------------
sage -t --warn-long 47.7 --random-seed=57011699320160764553253619770787736483 src/sage/plot/plot.py  # 1 doctest failed

[sagetrac-git]

Branch pushed to git repo; I updated commit sha1. This was a forced push. Last 10 new commits:

2b61474	spkg primecountpy
687687b	use primecountpy instead of built-in cython interface
9f6c4a1	remove primecount from various lists of optional spkgs
de7de85	cython is not in toolchain!
43d481b	bump to 0.1.0
f66821b	cysignals are needed
ae37b4e	deprecate sage.interfaces.primecount, not just remove
d3ae5f4	primecount is on conda, too
3a8c7fe	primesieve is on conda, too
74b3845	allow float inputs for prime_pi

[sagetrac-git]

Changed commit from 8ef586a to 74b3845

[dimpase]

comment:45

rebased an added ​74b3845 allow float inputs for prime_pi to deal with the need for prime_pi to accept floats as input.
All the relevant tests pass.

[mkoeppe]

comment:46

LGTM.

[mkoeppe]

comment:47

Marking this ticket "critical" in light of #33054 because it adds the conda package information

[vbraun]

Changed branch from u/dimpase/packages/primecountpy to 74b3845