------------------------------------------------------
lines 7-68 of file: at_cascade/csv/covariate_spline.py
------------------------------------------------------

{xrst_begin csv.covariate_spline}
{xrst_spell
  cov
}

Create Spline Functions For Covariates
######################################

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

covariate_table
***************
Is a ``list`` of ``dict`` representation of a
:ref:`csv.simulate@Input Files@covariate.csv` file.
All of the keys in this ``dict`` are covariate names except for
``node_name`` , ``sex``, ``age`` , ``time`` and ``omega`` .
All of the columns have been converted to type ``float`` except
for *node_name* and *sex* which have type ``str`` .

node_set
********
This is the set of nodes in :ref:`csv.simulate@Input Files@node.csv`

age_grid
********
This is a sorted list of float containing the age values
that are present for each node_name and sex.

time_grid
*********
This is a sorted list of float containing the time values
that are present for each node_name and sex.

spline_cov
**********
This is a dict of dict of dict.
Let spline be defined by

    for node_name in covariate_table
        for sex in covariate_table
            for cov_name a covariate name or omega
                spline = spline_cov[node_name][sex][cov_name]

Then spline is the corresponding function of age and time; i.e.,
{xrst_code py}
    z = spline(age, time)
{xrst_code}
sets *z* to the value of the covariate or omega
for the specified age, time, node_name, and sex where
*age*, *time*, and *z* are floats.

Side Effects
************
This routine aborts with an assert if covariate.csv does not have the
same rectangular grid in age and time for each (node_name, sex) pair.

{xrst_end csv.covariate_spline}