Particle-tracking-manager

What's Changed
* improved drift_model_config(), include PTM metadata with output file by kthyng in https://github.com/axiom-data-science/particle-tracking-manager/pull/24


**Full Changelog**: https://github.com/axiom-data-science/particle-tracking-manager/compare/v0.8.1...v0.8.2

What's Changed
* updating docs by kthyng in https://github.com/axiom-data-science/particle-tracking-manager/pull/22
* tons of updates to and for docs by kthyng in https://github.com/axiom-data-science/particle-tracking-manager/pull/23


**Full Changelog**: https://github.com/axiom-data-science/particle-tracking-manager/compare/v0.8.0...v0.8.1

What's Changed
* Some updates by kthyng in https://github.com/axiom-data-science/particle-tracking-manager/pull/21
* time_step_output behavior has changed — 1 hour by default
* time_step is now 5 min by default
* added Dcrit parameter for accurately finding where drifters are stranded in tidal flats
* vertical_mixing is True by default now
* added seafloor_action option
* fixed some Leeway/3D handling and log messaging
* export_variables are specific to drift_model as needed
* do not drop zeta anymore since used in opendrift
* output_file is now an option


**Full Changelog**: https://github.com/axiom-data-science/particle-tracking-manager/compare/v0.7.1...v0.8.0

What's Changed
* Updates by kthyng in https://github.com/axiom-data-science/particle-tracking-manager/pull/20


**Full Changelog**: https://github.com/axiom-data-science/particle-tracking-manager/compare/v0.7.0...v0.7.1

What's Changed
* Updates by kthyng in https://github.com/axiom-data-science/particle-tracking-manager/pull/19
* Now initialize all class attributes with None and removed usage of `hasattr` which simplifies and clarifies some code.
* Improved handling of `start_time`, `end_time`, `duration`, and `steps` in `manager.py` which fixed a bug in which users couldn't input `start_time` and have the simulation run successfully.
* simplified handling of `horizontal_diffusivity` in `opendrift` model.
* user can change `end_time`, `duration`, and `steps` and have the others update accordingly. Tests added to check this.
* changed known model "CIOFS_now" to "CIOFSOP" to avoid upper/lower issues and include "OP" for "operational".
* many more tests and improved behavior for attribute checks and updates


**Full Changelog**: https://github.com/axiom-data-science/particle-tracking-manager/compare/v0.6.0...v0.7.0

* is set up to tell `opendrift` ROMS reader to save the interpolator to a cache that is set up the first time it is run. This only works with the newest dev version of `opendrift` at the moment, and the files saved are hundreds of MB, but it speeds up the simulations pretty well (12 to 30 seconds).
* reworked which variables are dropped in which scenarios for `opendrift` and integrated with using wetdry vs static masks.
* added package `appdirs` to manage the cache for storing interpolator pickles.
* fix to CLI so duration input is formatted correctly.
* can now input `name` to accompany user-input `xarray Dataset` for `ocean_model`.
* added `ocean_model` "CIOFS_now" local and remote links.