NEP29/pandas/numpy

[graingert]

Is your feature request related to a problem? Please describe.
determining the EOL date of scientific python involves a little bit of 4D chess https://numpy.org/neps/nep-0029-deprecation_policy.html

Describe the solution you'd like
include the nep29/pandas/numpy release schedule in a nice color coded chart

Describe alternatives you've considered
paste the NEP29 doc directly into https://endoflife.date/nep29 and redirect https://endoflife.date/numpy and https://endoflife.date/pandas to there

Additional context
Add any other context or screenshots about the feature request here.

[welcome]

Thanks for opening your first issue here! Be sure to follow the issue template!

[graingert]

there's even a python script for it

from datetime import datetime, timedelta

data = """Jan 15, 2017: NumPy 1.12
Sep 13, 2015: Python 3.5
Dec 23, 2016: Python 3.6
Jun 27, 2018: Python 3.7
Jun 07, 2017: NumPy 1.13
Jan 06, 2018: NumPy 1.14
Jul 23, 2018: NumPy 1.15
Jan 13, 2019: NumPy 1.16
Jul 26, 2019: NumPy 1.17
Oct 14, 2019: Python 3.8
Dec 22, 2019: NumPy 1.18
Jun 20, 2020: NumPy 1.19
"""

releases = []

plus42 = timedelta(days=int(365*3.5 + 1))
plus24 = timedelta(days=int(365*2 + 1))

for line in data.splitlines():
    date, project_version = line.split(':')
    project, version = project_version.strip().split(' ')
    release = datetime.strptime(date, '%b %d, %Y')
    if project.lower() == 'numpy':
        drop = release + plus24
    else:
        drop = release + plus42
    releases.append((drop, project, version, release))

releases = sorted(releases, key=lambda x: x[0])


py_major,py_minor = sorted([int(x) for x in r[2].split('.')] for r in releases if r[1] == 'Python')[-1]
minpy = f"{py_major}.{py_minor+1}+"

num_major,num_minor = sorted([int(x) for x in r[2].split('.')] for r in releases if r[1] == 'NumPy')[-1]
minnum = f"{num_major}.{num_minor+1}+"

toprint_drop_dates = ['']
toprint_support_table = []
for d, p, v, r in releases[::-1]:
    df = d.strftime('%b %d, %Y')
    toprint_drop_dates.append(
        f'On {df} drop support for {p} {v} '
        f'(initially released on {r.strftime("%b %d, %Y")})')
    toprint_support_table.append(f'{df} {minpy:<6} {minnum:<5}')
    if p.lower() == 'numpy':
        minnum = v+'+'
    else:
        minpy = v+'+'
print("On next release, drop support for Python 3.5 (initially released on Sep 13, 2015)")
for e in toprint_drop_dates[-4::-1]:
    print(e)

print('============ ====== =====')
print('Date         Python NumPy')
print('------------ ------ -----')
for e in toprint_support_table[-4::-1]:
    print(e)
print('============ ====== =====')

[hugovk]

This sounds useful!

I think this should be at https://endoflife.date/nep29 rather than /numpy or /pandas because NumPy and Pandas and other Scientific Python are free to vary a bit from the actual NEP 29 policy, which defines a minimum to support and allows projects to freely drop the older ones at will (even if they're technically still supported upstream).

So would we have a table something like this?

Version	Released	NEP29 EOL
NumPy 1.14	Jan 06, 2018	Jan 07, 2020
Python 3.6	Dec 23, 2016	Jun 23, 2020
NumPy 1.15	Jul 23, 2018	Jul 23, 2020
NumPy 1.16	Jan 13, 2019	Jan 13, 2021
NumPy 1.17	Jul 26, 2019	Jul 26, 2021
NumPy 1.18	Dec 22, 2019	Dec 22, 2021
Python 3.7	Jun 27, 2018	Dec 26, 2021
NumPy 1.19	Jun 20, 2020	Jun 21, 2022
Python 3.8	Oct 14, 2019	Apr 14, 2023

[captn3m0]

Now that #1692 is merged, this is partially solved. I think it's better if we document each product individually (and be sure to mention NEP29, as in the current version.

The "4D-chess" is something we should be doing, and making it simple for end-users on a per-product basis.

[BiNZGi]

Closing this for inactivity.

[hugovk]

btw NEP 29 has been "superseded by the scientific python ecosystem coordination guideline SPEC 0 — Minimum Supported Versions".