--------------------------------------------------
lines 6-72 of file: at_cascade/get_database_dir.py
--------------------------------------------------

{xrst_begin get_database_dir}

Get Database Directory Corresponding To a Fit
#############################################

Prototype
*********
{xrst_literal ,
   # BEGIN_DEF, # END_DEF
   # BEGIN_RETURN, # END_RETURN
}

node_table
**********
is the node_table for this cascade as a ``list'' of ``dict``.
It can't be ``None``.

split_reference_table
*********************
is the :ref:`split_reference_table-name` as a ``list`` of ``dict``.
It can't be ``None``.
If the list has length zero,
we say that the table is empty.

node_split_set
**************
If :ref:`split_reference_table-name` is empty,
this argument must be None.
Otherwise it is a ``set`` of ``int`` containing the
:ref:`node_split_table@node_id` values that appear in the
node_spit table.

root_node_id
************
is the node_id for the :ref:`glossary@root_node`.

root_split_reference_id
***********************
If :ref:`split_reference_table-name` is empty,
this argument must be None.
Otherwise it is an ``int`` specifying the
:ref:`split_reference_table@split_reference_id`
that the root node corresponds to.

fit_node_id
***********
This argument is an ``int`` is the node_id for the  :ref:`glossary@fit_node`.

fit_split_reference_id
**********************
If :ref:`split_reference_table-name` is empty,
this argument must be None.
Otherwise it is an ``int`` specifying the
:ref:`split_reference_table@split_reference_id`
that the fit corresponds to.

database_dir
************
The return value is a ``str`` containing the directory,
relative to the :ref:`option_all_table@result_dir`,
where the database corresponding to the fit is (or will be) located.
In other words, the fit database has the following path:

   *result_dir*\ ``/`` *database_dir*\ ``/dismod.db``


{xrst_end get_database_dir}