Pyjams

* Changed to sphinx_book_theme for documentation.
* Correct all tests.
* Check if outin is Iterable even if inp is not in `array2input`.
* Split ncread and ncinfo in individual files.
* Rename readnetcdf file to ncread.
* Check if scalar is number in array2input.
* license_file deprecated in setuptools, use license_files.

* Set maximum number of function evaluations to N*log(N) formula in `sce`.
* Require keyword names after mask in `sce`.
* New call_func method in `sce` to catch new numpy behaviour that
single-valued array does not compare to scalar anymore. Same for sampling
a single random number.
* Change from np.ones*fill_value to np.full in `pack`.
* Add tests with pandas.Series and pandas.DataFrame to remaining functions.
* Changed from around to round in tests of the `air_humidity` module.
* Added _FillValue and missing_value to netcdf test files for `ncread` and
`ncio`.
* Cleaned tests for `ncread` and `ncinfo`.
* Raise error in Morris sampling if more than one component changed at once.

* `ncio.create_variables` sets missing_value attribute if present even if
used for _FillValue.
* `ncio.get_variable_definition` returns _FillValue if present and only
missing_value if _FillValue is not present.
* Add tests with pandas.Series and pandas.DataFrame to (almost) all
existing functions.
* Remove from future imports from all ``pyjams`` main modules (not jams)
and all tests.
* GUI dialog functions of module `fgui` only imported if Tkinter installed.
* Support pandas.Series in `argmin`, `argmax`, and `argsort`.
* Support pandas Series and DataFrame in `closest` and `mad`.
* Bug in `esat` if multi-dimensional array: wrong array index.
* Support pandas.Series and pandas.DataFrame in helper functions
`input2array` and `array2input`.
* Add pandas as requirement for ``pyjams``.

* Merged old library JAMS (https://github.com/mcuntz/jams_python) as a
subdirectory. Remove deprecated routines that are in the main part now.
Remove unsupported routines.

* Set filename without suffix as default plot name in `mcPlot`.
* Exit if initial population failed twice in `sce`.
* Allow ncol=1 in `color.get_cmap`.
* prepend and append keywords in `mad`.
* Warn only if lb > ub, simply set mask if lb == ub in `sce`.
* Add functions `pack`, `unpack` mimicking Fortran intrinsics.

* Replace plotly with hvplot in `mcPlot`.
* Add unicode symbol degree \u00B0, which gets replaced by ^\circ
if usetex==True in str2tex.
* Do not escape % if not usetex in str2tex.
* Make `get_variable_definition` public in ncio.
* Correct docstring of strip keyword in fsread and xread.
* Added `filebase` to helpers.
* Exchanged functions and wrapper functions in readnetcdf.