Dtscalibration

------------------
Bug fixes

* Fixed deprecated np.float and np.int

Other

* Moved project build system & scripts to hatch.
* Project is now fully configured using pyproject.toml
* Linting is handled by Ruff.
* Notebooks are now rendered as part of the documentation.

Removed

* cli.py script (unused) has been removed.

New contributors:
* Karl Lapo (klapo)

------------------

* Added support for Sensornet Sentinel v5 files.

------------------

* Deprecated Python 3.7
* Fixed documention rendering
* Updated readme badges

------------------
New features

* Added support for Python 3.9, 3.10.
* Silixa's xml version 8 is now supported

Bug fixes

* Loading in untested sensornet files will not give a UnboundLocalError error anymore
* Sensornet .ddf file version check is now more robust (commas are replaced to periods)
* Changed matplotlib's deprecated DivergingNorm to TwoSlopeNorm
* Updated the stokes_variance_linear docstring to remove incorrect and duplicate information
* Adjusted resample_datastore to avoid using deprecated 'base' kwarg, instead using the new arguments 'origin' and 'offset'. See http://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.resample.html

Others

* Notebook 04 on Stokes variance has been updated to explain the different estimation methods for the variance, and their importance.
* Notebook 07 and 08 have been updated to take into account the changes in ds.stokes_variance.
* Silenced xarray's __slots__ warning
* Deprecated Python 3.6
* CI has been moved to GitHub Actions. Travis-CI and AppVeyor have been removed.

------------------
* Same as v1.0.1

------------------
New features

* st_var can now be array-like. The array-like can vary in x or time, or both.
* When converting from a xr.Dataset to a DataStore object, the attributes arenow transfered over
* Added 'verbose' kwarg to shift_double_ended utility function to silence theconsole output

Bug fixes

* If the '_sections' attribute is missing from a DataStore object it is automagically re-added.
* Assert that stokes variances are passed when running a double-ended WLS calibration
* Added check for NaN/inf values in wls_sparse solver to avoid unclear error messages
* Before calibration, the values of the used reference temperatures are checked if they are valid (float dtype, no NaN/inf values, correct time dimension)

Others

* European Geosciences Union conference 2020 presentation added
* Updated documentation with content article
* Use YAPF formatting of the Python files
* Travis-CI.org integration in GitHub restored.