Fmskill

* It is now possible to provide a track ModelResult as a pandas DataFrame

* DHI Altimetry data download from api
* Hit ratio metric
* Custom metric example
* Add more modelresults to existing comparer
* Better merging of comparers also if time period is overlapping
* Fix bug in TrackComparer

* ModelResult now supports NetCDF including multiple-file NetCDF data using xarray as backend
* An example can be seen here: https://nbviewer.jupyter.org/github/DHI/fmskill/blob/main/notebooks/NetCDF_ModelResult.ipynb
* ModelResult also supports xarray input
* Connector can now write a config file with information about filenames, items etc of ModelResults and Observations
* Connector can write config to dict or to yaml or xlsx files
* Connector can also read config files from both yaml and xlsx files

Re-work on how observations and modelresults are connected. Observations are no longer "added" to the modelresult. With this release, a new object "Connector" has been introduced. It is responsible for connecting observations and modelresults, including validation and extraction of modelresult data at observation positions.

Note! This release has breaking changes!

* Taylor diagram
* Quick and dirty compare by fmskill.compare()
* Bugfixes

Skill object with better visualization possibilities for skill assessments. The main methods of the new class are:
* style()
* plot_bar()
* plot_line()
* plot_grid()
* sel()