Update internal API

work with modern Pandas. Simplify internals. This is an interim release till I can implement the PR with new data. Thanks for your patience.

improve rolling average computation

this avoids contaminating many days rolling average with NaN from a single bad data point, using pandas.rolling.mean()

new source of recent Ap, f10.7 data, Python >= 3.7

SWPC obseleted their source of recent data. As a result, we have to get Ap and f10.7 from two discrete web sources. This was a long-term pain point that has been resolved in this release.
Part of the release also needs Python >= 3.7 to make this package reliable across Python install mechanisms, including venv and non-CPython installs.

src/ layout, __main__ instead of scripts

src/ layout:

'

BUGFIX: Kp value scaling

• BUG: fix Kp to have proper scaling
• move CI to github actions
• always do "install" tests

working ok in virtualenv as prereq

include_package_data=True seems to be the last piece needed to work in all situations as an installed prereq, including virtualenv

use pkg_resources for data/ path

To help virtualenv use such as Travis-CI, use pkg_resources to resolve data_files data/ path. zip_safe: false was also required in setup.cfg

data path

default data path data/

vendorize yeardec2datetime, more robust CI, pass out time resolution

parametrize tests, download current year if full year not yet available

corner case debug

corner caes

corner

pass along time resolution

initial release

Following our design pattern of "one tool for one job" in Python, we provide geomag-indices package, which we use in numerous public and proprietary programs.