Pygaps

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

New features:

* 🐍 Minimum python reduced to 3.6 to increase compatibility.
* 🎆 pyGAPS is now (actually) on ``conda-forge``.

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

New features:

* ⚠️ To reduce import bloat pyGAPS is now **modular**. Previously global
functions must be accessed through submodules. For example:
``pygaps.area_BET`` now must be imported from
``pygaps.characterisation.area_BET``.
* 🎆 pyGAPS is now on ``conda-forge``
* Volumetric adsorbed amount is now given in either ``volume_gas`` or
`volume_liquid` basis, corresponding to the volume amount the adsorbate would
occupy in the bulk gas phase, or the volume of an ideal liquid phase of
adsorbate (at isotherm temperature). The old loading_basis of ``volume`` is
deprecated and automatically converted to ``volume_gas`` while emitting a
warning.
* Manufacturer parsing is streamlined with the introduction of the
``pygaps.parsing.isotherm_from_commercial`` function.
* Drastically improved reliability and modularity of code in preparation
for the release of the pyGAPS-gui interface.
* Added a "zero" thickness model that assumes no mono/multilayer sorption.
Useful in the case of condensation in hydrophobic materials.
* Better documentation.

Changes:

* ⚠️ Minimum python is now 3.7, maximum increased to 3.10.
* Removed the need to pass DataFrame column names with the ``other_keys``
syntax. PointIsotherms, now save *all* passed DataFrame columns.
* Smart assigning of isotherm metadata caused confusion and was removed.
Metadata assigned like ``isotherm.myparam`` is now no longer serialized to
dictionaries, parsers etc. Instead, isotherms have a `isotherm.properties`
dictionary which contains all metadata.
* Adsorption/desorption branches are internally represented as 0 and 1 instead
of False/True. This allows the possibility for further cycles to be introduced
in a future release.
* Isotherm ``material`` and ``adsorbate`` are now always instantiated as
``pygaps.Material`` and ``pygaps.Adsorbate`` classes.

Plus many small and large bugs fixed.

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

New features:

* pyGAPS is now capable of parsing to and from Adsorption Information Files
(AIF). For more info, see *Evans, Jack D., Volodymyr Bon, Irena Senkovska, and
Stefan Kaskel. ‘A Universal Standard Archive File for Adsorption Data’.
Langmuir, 2 April 2021, DOI: 10.1021/acs.langmuir.1c00122.*

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

New features:

* The internal Adsorbate list now contains over 170 analytes, 81 of which have a
correspondence in the thermodynamic backend. This includes multiple vapours,
VOCs, and refrigerants.
* Added new mode for pressure as "relative%", which represents relative pressure
as a percentage rather than a fraction (i.e. p/p0 * 100).
* Added two new modes for loading as "fraction" and "percent". This ties the
uptake to the same basis as the adsorbate (i.e. weight%, volume%, mol% or
fractions thereof).
* Conversely, NIST format isotherms based on "wt%" can also be converted.
* Significant improvements to the Horvath-Kawazoe methods of pore size
distribution, including more pore geometries (cylindrical and spherical
through the Saito-Foley and Cheng-Yang modifications) and the inclusion of
extended HK models, with the Cheng-Yang and Rege-Yang corrections.
* Command-line interface: a CLI entry point is automatically added during pyGAPS
installation and can be called with ``pygaps -h`` to perform some simple
commands.
* Isotherm JSON parser (``pygaps.isotherm_to_json``) now passes all extra
parameters to the ``json.dump`` function.
* Perform isotherm branch separation based on maximum pressure, rather than
first derivative. In such way, slight uncertainty in pressures won't lead to a
wrong assignment.
* The reference area for an alpha_s calculation can now be specified as either
"BET" or "Langmuir".
* Convenience function `isotherm.convert()` which combines all isotherm
conversion modes.
* Convenience function `pygaps.model_iso()` which fits a model to a
PointIsotherm, effectively wrapping `ModelIsotherm.from_pointisotherm`.
* Convenience functions for isotherm parsing: `isotherm.to_json()`,
`isotherm.to_csv()` and `isotherm.to_xl()`.
* Parsing from instrument output files now gets more information.
* Plot quality has been overall improved.
* Improved performance for isotherm conversions.
* General refactoring and speed-ups.
* Switched to GitHub actions for CI, now MacOS builds are also tested.

Breaking changes:

* Included Python 3.8 and deprecated Python 3.5.
* All parameters like ``adsorbate_basis`` or ``adsorbate_unit`` have been
changed to ``material_basis`` and ``material_unit`` for consistency. Old
format should still work for some time.
* Some model names have been changed to include only ASCII: ``JensenSeaton``,
``FHVST``, ``WVST``.
* For isotherm pressure/loading, a `limits` tuple is now passed instead of
`min_range` and `max_range`, as for other functions in pyGAPS.
* JSON ModelIsotherms now have ``name`` instead of ``model`` as the model name.
This is now consistent with both CSV and Excel.
* The `isotherm_to_jsonf` and `isotherm_from_jsonf` functions have been removed.
Functionality has been merged with `isotherm_to_json` similarly to the `pandas
model
<https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.read_json.html>`_.
* Removed the `util_get_file_paths` function.

Fixes:

* Volumetric -> molar conversions were not calculated correctly.
* Isosteric enthalpy could not be calculated if the isotherm was not in mmol/g.
* ModelIsotherm creation could in some cases ignore isotherm branch splitting.
* BET area now attempts to pick at least 3 points if physically consistent.
Should stop failing on some isotherms.
* BET/Langmuir area maximum calculation was offset by one point.
* The "section" returned in tplot/alphas is now consistent for both manual and
automatic limits: a list indices for selected points

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

New features:

* Added fluids to database: n-pentane, n-hexane, n-octane, o-xylene, m-xylene,
p-xylene, cyclohexane, hydrogen sulphide and sulphur hexafluoride.

Fixes:

* Converting Adsorbates to a dictionary now correctly outputs the list of
aliases.
* Changed stored critical point molar mass values for some adsorbates.

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

* Fixed error in dft kernel acquisition.
* Removed duplicate plot generation from virial initial henry.
* Fixed Appveyor testing.