Aurora-asi-lib

Changed
- Removed support for python 3.8
- Added support for python 3.12
- Incremented two package versions in requirements.txt: `scipy==1.20.0` and `h5py==3.10.0`.

Added
- Project metadata in `pyproject.toml` and removed `setup.cfg`.
- Additional package URLs in PyPI.

Fixed
- Auroral intensities resulted in an index error is the satellite was at the horizon.

Changed
- Removed the RGB normalization in the `trex_rgb()` loader. This fixed the vertical stripes in the keograms, but made the fisheye and mapped images much darker (since the `norm` kwarg in `plt.pcolormesh` and `plt.imshow` does nothing).
- Refactored the TREx and Imager tests reflecting the minor changes.

Added
- A `color_brighten` kwarg to by default enhance the RGB colors when calling the following asilib.Imager methods, `plot_fisheye`, `animate_fisheye_gen`, `plot_map`, and `animate_map_gen`. Unless `color_brighten=False`, the plots remain the same.

Fixed
- A bug raised in issue 15 (https://github.com/mshumko/aurora-asi-lib/issues/15) where an `asilib.Imagers` class, initiated with a single `asilib.Imager`, would try to incorrectly index `asilib.Imager` unless it is wrapper in a tuple.

Added
- The `n_avg` and `imager` kawrgs to the `psa_emccd()` function. These kwargs allow for custom Imager instance, as well as average the images over `n_avg` times.

Added
- `custom_alt` kwarg to the THEMIS, REGO, and TREx loaders. Credit: Cassandra M.
- A test for the `custom_alt` functionality.
- Tests for the custom colors.

Changed
- How RGB color channels are loaded. By picking one or multiple color channels, the underlying data for the unselected channels is masked as NaNs. matplotlib handles these images well.