Note that Numpy is still a dependency of Skyfield and its dependencies.

* chore: fix files to commit on tag

* fix(ephemerides): fix a bug that made the ephemerides calculations impossible for the Poles (#21)

* chore: fix the target branch on Dependabot's config

* ci: remove Commitlint workflow (replaced by semantic-pull-request) (#25)

* ci: remove Commitlint workflow (replaced by semantic-pull-request)

* ci: fix doctests not running correctly on some OS and Python versions

* Add season change event in events.py and enum.py

* Add minor change in season change event

* Add documentation, change enum constants, change output object of _search_earth_season_change

* Minor changes to _search_earth_season_change

* Update EventType enum, update class of details

* Fix minor bugs

* docs: Update docs for _search_earth_season_change and _search_conjunction.chore: make minor changes to _search_earth_season_change

* Update: minor changes to match Python coding style.

* Update: minor changes to match Python coding style. Docs: update docstring of _search_earth_season_change and _search_conjunction

* test:update legacy tests for events.py. update: update enum.py and events.py to match black coding style.

* Fix some minor issues with Black and Event.details field

Co-authored-by: Jérôme Deuchnord <jerome@deuchnord.fr>
Co-authored-by: Jérôme Deuchnord <Deuchnord@users.noreply.github.com>

BREAKING CHANGE: the `Event.details` field is now a dictionary (was previously a string).

BREAKING CHANGE: name of objects of type `kosmorrolib.model.Object` has
been replaced with an enum typed identifier.

BREAKING CHANGE: `kosmorrolib.data` has been renamed to
`kosmorrolib.model`. To ensure further BC-break to happen on this side,
prefer using the model now exposed from kosmorrolib directly.

BREAKING CHANGE: name of objects of type `kosmorrolib.model.Object` has
been replaced with an enum typed identifier.

BREAKING CHANGE: `kosmorrolib.data` has been renamed to
`kosmorrolib.model`. To ensure further BC-break to happen on this side,
prefer using the model now exposed from kosmorrolib directly.