Update distribution pipeline

[psomhorst]

This PR started out small, but I added some improvements to several parts of the distribution pipeline.

• To ensure eitprocessing keeps working on newer versions of Python, I added the test and build action for Python 3.11-3.13.
• The image containing the data (ghcr.io/eit-alive/eittestdata:latest) only has Python 3.10 with pip installed. Instead of editing that image, I edited the GitHub action to mount the data from the image, and run the test and build on python:3.x-bookworm instead. This makes it easier to adapt to new Python versions and replace the data image in the future, if necessary. See comment below for new description.
• The coveralls action was not working. I removed it from the test action and created a new action. Coveralls does not report well when run on pull requests (and they don't really make sense), so I restricted it to running on pushes to main and develop.
• I updated the linting action to install the ruff version from pyproject.toml. I already forgot to update the version in the action once, not planning on doing it again.
• I updated the classifiers in the pyproject.toml file to better represent the (current status of) the project.
• I added a py.typed file to the root of the package. This enables linters to use the source code for type checking when installing eitprocessing from package. (See https://peps.python.org/pep-0561/#packaging-type-information and https://mypy.readthedocs.io/en/stable/installed_packages.html#creating-pep-561-compatible-packages.)

[DaniBodor]

This looks fine to me at first glance, but I am not very fluent at docker, so asked @wbaccinelli to review it instead. Also, things will change again soon in #303

[DaniBodor]

maybe also add the other authors above here, to make it match the CFF file

[psomhorst]

This looks fine to me at first glance, but I am not very fluent at docker, so asked @wbaccinelli to review it instead. Also, things will change again soon in #303

I wasn't aware of #303. Maybe it's best to do that in this PR, instead of changing it twice. I think we can already remove the container requirement by extracting the data files from the container (https://github.com/marketplace/actions/docker-extract). This can then later be replaced, e.g. by downloading from Zenodo (although downloading from Zenodo seems to be quite slow).

I'll look into this tomorrow.