Rayflare

------------------------------------
**Highlights:**

- **Speed improvements for all ray-tracing calculations**, by using numba jit (just-in-time) compilation of heavily-used
functions, and more efficient lookup of TMM values for integrated ray-tracing/TMM calculations. The use of just-in-time
compilation can be turned on or off, see here: :ref:`Performance <performance>`.
- **New analytical method for ray-tracing**, which is much faster than full ray-tracing and can be used with full
accuracy for surfaces where the number of ray interactions is known exactly in advance (e.g. upright pyramids with
opening angles between 45 and 54 degrees). Note that the calculation of absorption profiles in interface layers
is not yet implemented for this method (though this is on the roadmap, as there is no fundamental reason this
cannot be done). Absorption in bulk layers is calculated as usual if the analytical method is used. This option
can be toggled on per surface in an rt_structure object, see :ref:`here <textures>`. The analytical method is not
yet implemented for calculating redistribution matrices (for the angular redistribution method), but this is
planned for the next minor release.
- **Phong scattering**: a new option for the ray-tracer, which allows additional non-specular scattering. The continuum
between perfectly specular and Lambertian scattering can be controlled with the Phong exponent; see the documentation
:ref:`here <textures>`.
- **Rigorous treatment of polarization in ray-tracing** (no changes to treatment of polarization in TMM and RCWA calculations).
See further mathematical details of the new method :ref:`here <polarization>`.
- Option to set a maximum number of bulk passes for the ray-tracer using the maximum_passes option
(see :ref:`Options <options>`).
- More detailed output from ray-tracing calculations. See the docstring for :literal:`rt_structure.calculate`
:ref:`here <ray_tracing>`.
- More detailed output from RCWA calculations.

**Possible backwards compatibility issues:**

- The (previously undocumented) returns from rt_structure.calculate with details on final ray directions and intensities
have been reorganized. See the docstring :ref:`here <ray_tracing>`.
- Previously, the ray-tracing method did not handle polarization correctly, assuming a constant mixture
of *s* and *p* polarization throughout the ray-tracing procedure regardless of the angle of the surface
textures. Thus, ray-tracing
calculations may now give slightly different results. For most cases, especially for initially
unpolarized light, the differences are small, but this
depends on the surface texture, layers, and materials used.

------------------------------------
**Highlights:**

- New RCWA method (`Inkstone <https://github.com/alexysong/inkstone>`_) now integrated with RayFlare.
This is an all-Python program which is therefore easy to install (unlike S4). This means all core u
functionality of RayFlare is now available without the need to compile anything on your computer.
Users can toggle between S4 and Inkstone by setting the ``RCWA_method`` option.
- Calculations with unpolarized light using RCWA should be faster now.
- The ``wavelengths`` user options is now called ``wavelength``, for compatibility with Solcore.
``wavelengths`` still works, but will be deprecated in future.

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

**Highlights:**

- New functionality with ray-tracing, useful for e.g.
`perovskite/Si <https://rayflare.readthedocs.io/en/latest/Examples/perovskite_Si_rt.html>`_ or III-V/Si tandems. \
- New functions to generate textures: rough pyramids, rough planar surfaces, and (rough or smooth) hemispherical caps.
See :ref:`here <textures>` for further details.


**Possible backwards compatibility issues:**

- The :literal:`make_absorption_function` function now requires a slightly different input format (the full result
of the optical calculation, rather than just the profile part). This is to standardize its behaviour across all
methods. See :ref:`the documentation <utilities>` for more details.


This is a major release with a lot of new features and fixes. The biggest change is the introduction
of functionality for ray-tracing simulations: previously, it was possible to use ray-tracing in conjunction with the
transfer-matrix model to calculate reflection/transmission/absorption probabilities at interfaces only when using the
angular redistribution method. It is now possible to do this in a pure ray-tracing simulation. Generally, if you do not
need to use another method for one of the interfaces, doing a pure ray-tracing simulation is faster than using the
angular redistribution matrix method with ray-tracing. An example of the new simulation, which is useful for modelling
e.g. perovskite on Si or III-V on Si solar cells with textured silicon surfaces, can be found
`here <https://rayflare.readthedocs.io/en/latest/Examples/perovskite_Si_rt.html>`_.

It is now also possible to specify that previous results for the angular redistribution matrices should be overwritten
rather than loaded. This is useful if you have changed the geometry of the system and want to recalculate without changing
the project name or manually deleting files every time. This can be done by setting the :literal:`overwrite` keyword
argument to :literal:`True` when calling :literal:`process_structure`. See :ref:`the docstring <matrix>` here.

There were also two bugs relating to the calculation of the absorption profile, both of which were apparent only in
very specific circumstances which are unlikely to occur in a real solar cell (highly transparent absorbers). The details
of what was changed and what was wrong previously are explained
`here <https://github.com/qpv-research-group/rayflare/issues/56>`_.

.. _mailing list: https://www.solcore.solar/mailing-list