SPEC 0 — Minimum Supported Versions

Description

This SPEC recommends that all projects across the Scientific Python ecosystem adopt a common time-based policy for dropping support of older Python and core package versions.

All versions refer to feature releases (i.e., Python 3.8.0, NumPy 1.19.0; not Python 3.8.1, NumPy 1.19.2).

Specifically, we recommend that:

1. Support for a given version of Python be dropped 3 years after its initial release.
2. Support for a given version of other core packages be dropped 2 years after their initial release.

Motivation

Limiting the scope of supported dependencies is an effective way for packages to limit maintenance burden. Combinations of packages need to be tested, which impacts also on continuous integration times and infrastructure upkeep. Code itself also becomes more complicated when it has to be aware of various combinations of configurations.

Adoption of this SPEC will ensure a consistent support policy across packages, and reduce the need for individual projects to divise similar policies.

Ultimately, reduced maintenance burden frees up developer time, which translates into more features, bugfixes, and optimizations for users.

Background

In the past, longer support cycles were common. There were several reasons for this, including the Python 2 / 3 transition, difficulties installing packages, and users needing to use old, operating-system provided versions of Python. The situation has since improved due to improved installations via binary wheels, virtual environments becoming commonplace, and support for Python 2 being dropped.

Drop Schedule

On Oct 13, 2022 drop support for python 3.8 (initially released on Oct 14, 2019)
On Dec 15, 2022 drop support for scikit-image 0.18.0 (initially released on Dec 15, 2020)
On Dec 22, 2022 drop support for scikit-learn 0.24.0 (initially released on Dec 22, 2020)
On Dec 26, 2022 drop support for pandas 1.2.0 (initially released on Dec 26, 2020)
On Dec 31, 2022 drop support for scipy 1.6.0 (initially released on Dec 31, 2020)
On Jan 30, 2023 drop support for numpy 1.20.0 (initially released on Jan 30, 2021)
On Mar 26, 2023 drop support for matplotlib 3.4.0 (initially released on Mar 26, 2021)
On Jun 20, 2023 drop support for scipy 1.7.0 (initially released on Jun 20, 2021)
On Jun 22, 2023 drop support for numpy 1.21.0 (initially released on Jun 22, 2021)
On Jul 02, 2023 drop support for pandas 1.3.0 (initially released on Jul 02, 2021)
On Jul 08, 2023 drop support for networkx 2.6 (initially released on Jul 08, 2021)
On Sep 24, 2023 drop support for scikit-learn 1.0 (initially released on Sep 24, 2021)
On Oct 05, 2023 drop support for python 3.9 (initially released on Oct 05, 2020)
On Nov 16, 2023 drop support for matplotlib 3.5.0 (initially released on Nov 16, 2021)
On Dec 03, 2023 drop support for scikit-image 0.19.0 (initially released on Dec 03, 2021)
On Dec 31, 2023 drop support for numpy 1.22.0 (initially released on Dec 31, 2021)
On Jan 22, 2024 drop support for pandas 1.4.0 (initially released on Jan 22, 2022)
On Feb 05, 2024 drop support for scipy 1.8.0 (initially released on Feb 05, 2022)
On Feb 28, 2024 drop support for networkx 2.7 (initially released on Feb 28, 2022)
On Apr 08, 2024 drop support for networkx 2.8 (initially released on Apr 09, 2022)
On May 11, 2024 drop support for scikit-learn 1.1.0 (initially released on May 12, 2022)
On Jun 21, 2024 drop support for numpy 1.23.0 (initially released on Jun 22, 2022)
On Jul 28, 2024 drop support for scipy 1.9.0 (initially released on Jul 29, 2022)
On Oct 03, 2024 drop support for python 3.10 (initially released on Oct 04, 2021)

Implementation

Core Project Endorsement

Ecosystem Adoption

Notes

• This document builds on NEP 29, which describes several alternatives including ad hoc version support, all CPython supported versions, default version on Linux distribution, N minor versions of Python, and time window from the X.Y.1 Python release.

• Code to generate support and drop schedule tables:

from datetime import datetime, timedelta
from glob import glob
import json
import requests
from packaging.version import parse, Version


py_releases = {
    # "3.7": "Jun 27, 2018",
    "3.8": "Oct 14, 2019",
    "3.9": "Oct 5, 2020",
    "3.10": "Oct 4, 2021",
}
core_packages = [
    # Path(x).stem for x in glob("../core-projects/*.md") if "_index" not in x
    "numpy",
    "scipy",
    "matplotlib",
    "pandas",
    "scikit-image",
    "networkx",
    "scikit-learn",
]
plus36 = timedelta(days=int(365 * 3))
plus24 = timedelta(days=int(365 * 2))

# Release data


def get_release_dates(package, support_time=plus24):
    releases = {}
    response = requests.get(f"https://pypi.org/pypi/{package}/json").json()

    for ver, ver_files in response["releases"].items():
        try:
            version = Version(ver)
        except:
            continue

        if version.is_prerelease or version.micro != 0:
            continue

        upload_time = min(
            (release_file["upload_time_iso_8601"] for release_file in ver_files),
            default=None,
        )
        if upload_time is None:
            continue

        release_date = None
        for format in ["%Y-%m-%dT%H:%M:%S.%fZ", "%Y-%m-%dT%H:%M:%SZ"]:
            try:
                release_date = datetime.strptime(upload_time, format)
            except:
                pass

        if not release_date:
            continue

        drop_date = release_date + support_time
        if drop_date >= datetime.now():
            releases[ver] = {
                "release_date": release_date,
                "drop_date": drop_date,
            }

    return releases


# {
#    "python": {
#        "3.8": {
#            "release_date": datetime.datetime(2019, 10, 14, 0, 0),
#            "drop_date": datetime.datetime(2022, 10, 13, 0, 0),
#        },
#        ...
#    },
#    "numpy": {
#        ...
#    },
#    ...
# }
package_releases = {
    "python": {
        version: {
            "release_date": datetime.strptime(release_date, "%b %d, %Y"),
            "drop_date": datetime.strptime(release_date, "%b %d, %Y") + plus36,
        }
        for version, release_date in py_releases.items()
    }
}
package_releases |= {package: get_release_dates(package) for package in core_packages}


# Print Gantt chart

print(
    """<!-- prettier-ignore-start -->



  
gantt
dateFormat  YYYY-MM-DD
axisFormat  %m / %Y
title Support Window
"""
)

for name, releases in package_releases.items():
    print(f"\nsection {name}")
    for version, dates in releases.items():
        print(
            f"{version}  :     {dates['release_date'].strftime('%Y-%m-%d')},{dates['drop_date'].strftime('%Y-%m-%d')}"
        )

print("

\n<!-- prettier-ignore-end -->")

# Print drop schedule

rel = {}
for name, releases in package_releases.items():
    rel |= {
        " ".join([name, str(ver)]): [dates["release_date"], dates["drop_date"]]
        for ver, dates in releases.items()
    }


rel = dict(sorted(rel.items(), key=lambda item: item[1][1]))
for package, dates in rel.items():
    print(
        f"On {dates[1].strftime('%b %d, %Y')} drop support for {package} "
        f"(initially released on {dates[0].strftime('%b %d, %Y')})"
    )