Known issues
============

Multiprocessing
---------------

.. Warning::
   Make sure to include the :code:`if __name__ == "__main__":` guard in the script being run.
   Omitting this can cause issues when running in parallel mode, specifically on Windows.

On some systems parallelisation using `multiprocessing`_ doesn't work properly.
To fix this, you can run in single-thread mode, by passing
:code:`parallel=False` to the run method. For example:

.. code-block:: python

   from drama import oscar

   results = oscar.run(semiMajorAxis=[6500, 6550], parallel=False)

This can also help to debug errors, as the traceback becomes more readable.

.. _multiprocessing: https://docs.python.org/3.12/library/multiprocessing.html

Lists in default config
-----------------------

Some config entries are required to always be a list (such as :code:`populationCloudIds` in :func:`drama.ares.get_basic_config`).
These inputs currently can't be parsed correctly, and cause an error when passed to the `run` and `run_lazy` functions.
As a workaround, the config dictionary can be passed as a list:

.. code-block:: python

   from drama import ares

   config = ares.get_basic_config()
   results = ares.run([config])

This requires manually generating a list of configs to do multiple runs, when these config entries need to be included.
This would look like the following:

.. code-block:: python

   from drama import midas

   config1 = {'sourceSwitches': [0, 0, 0, 0, 0, 0, 0, 0, 0, 1]} # Condensed population only
   config2 = {'sourceSwitches': [1, 1, 1, 0, 0, 0, 0, 0, 0, 0]} # Fragments/LMRO only
   results = midas.run([config1, config2])

If running a parametric analysis, make sure the entries which are lists by default are not included,
or have been removed when using the default config as a starting point:

.. code-block:: python

   from drama import ares

   config = ares.get_basic_config()
   # delete keys containing lists
   keys = [
      "populationCloudIds",
      "populationCloudSwitches",
      "radarEquationParameters",
      "userEnteredCovariances",
      "uncertaintyScalingFactors",
      "ACPLValues",
      "revolutionVectors",
   ]
   for key in keys:
      config.pop(key)
   # add lists for parametric runs
   config['semiMajorAxis'] = [sma for sma in range(6800, 7200, 100)]
   results = ares.run(config)

This will use the default values, or the values set in the project file if used, instead.
The following config entries are affected by this issue:

.. code-block:: python

   ares_keys = [
      "populationCloudIds",
      "populationCloudSwitches",
      "radarEquationParameters",
      "userEnteredCovariances",
      "uncertaintyScalingFactors",
      "ACPLValues",
      "revolutionVectors",
   ]
   midas_keys = [
      "sourceSwitches",
      "backgroundMeteoroidPopulation",
      "cloudSwitches",
      "cloudFileIds",
      "radarEquationParameters",
   ]
   oscar_keys = [
      "generatePlots",
   ]