.. _processing-pipeline:

Processing Pipeline
===================

Post-process Grids
------------------

After the MESA grids have finished, the important data needs to be extracted
from them. This is a multistep process that we have automated as much as possible.
Central in this process is the ``PSyGrid`` object, which is used to read and
store the MESA data. You can learn more about it below.

In general, the post-processing consists of the following steps:

1. ``step_1``: Extracts the important data from the MESA output files and stores it in a
   ``PSyGrid`` object.
2. ``step_2``: Combine multiple ``PSyGrid`` objects into a single one.
3. ``step_3``: Add additional information to the ``PSyGrid`` object, such as
   supernova outcomes and remnant masses.
4. ``step_4``: Train classifiers and initial-final interpolators based on the
   ``PSyGrid`` data.

After completing these steps, the ``PSyGrid`` object contains all the
necessary information to perform population synthesis simulations.
The following diagram illustrates the processing pipeline:

.. image:: ../_static/processing_pipeline.png
   :alt: Processing Pipeline
   :align: center
   :width: 100%

As part of the post-processing, we also generate plots and tables to
summarize the results. Furthermore, the post-processing also allows for the
selection of reruns, which can be used to rerun a specific subset of the
initial grid with a different numerical MESA setup to reach convergence.

The post-processing steps can be run individually or as part of an
automated pipeline (the ``ini`` file configuration).
The following sections provide more details on each step and the available tools.


``PSyGrid`` object
~~~~~~~~~~~~~~~~~~

.. toctree::

   post_processing/psygrid

Our post-processing pipeline
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

.. toctree::

   post_processing/pipeline

:samp:`ini` file documentation
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Learn how to use an :samp:`ini` file to set up the post-processing pipeline.

.. toctree::

   post_processing/pipeline_ini