-------------------------------------------
lines 5-105 of file: example/user/csv2db.py
-------------------------------------------

# {xrst_begin user_csv2db.py}
# {xrst_spell
#     mtall
# }
# {xrst_comment_ch #}
#
# csv2db_command: Example and Test
# ################################
#
# Using This Example
# ******************
# See :ref:`user_example@Run One Example` for instructions
# on running just this example.
# After doing that, one can run the command
# ::
#
#     python/bin/dismodat.py build/example/user/example.db
#
# To generate the csv files corresponding to the example database.
# One can then inspect the csv files in the ``build/example/user``
# to see all the relevant information.
#
# Discussion
# **********
# The following describes the mode and data for this example:
#
# rate_true
# *********
# The true value for the rates, used to simulate the data,
# are constant w.r.t age and time and are given by:
# {xrst_literal
#     BEGIN RATE_TRUE
#     END RATE_TRUE
# }
#
# P
# *
# The notation :math:`P` is used for prevalence.
# The initial prevalence at age zero is zero; i.e. :math:`P(0) = 0`.
# We use :math:`S(a)` (:math:`C(a)`) for the susceptible
# (with condition) fraction of the initial population.
# The true prevalence :math:`P(a) = C(a) / [S(a) + C(a)]`
# is solved for using the ODE:
#
# .. math::
#  :nowrap:
#
#  \begin{eqnarray}
#  S(0)    & = & 1 \\
#  C(0)    & = & 0 \\
#  S'(a)   & = & - \iota S(a) + \rho C(a)  - \omega S(a) \\
#  C'(a)   & = & + \iota S(a) - \rho C(a)  - \omega C(a) - \chi C(a)
#  \end{eqnarray}
#
# Rate Grids
# **********
# The value of omega is modeled as know and equal to the
# value of
# :ref:`csv2db_command@integrand@mtall`
# corresponding to the age-time intervals:
# {xrst_literal
#     BEGIN INTERVALS
#     END INTERVALS
# }
# The non-zero rates (iota, rho, chi) are modeled as unknown and piecewise
# bilinear with the same grid points.
#
# Data
# ****
# The Data is simulated,
# without any noise, for the following integrands:
# :ref:`csv2db_command@integrand@remission` ,
# :ref:`csv2db_command@integrand@mtexcess` ,
# :ref:`csv2db_command@integrand@prevalence` .
# Note that the model for the noise in the measurement
# :ref:`csv2db_command@meas_std`
# is a 10 percent coefficient of variation.
# See the file
# :ref:`db2csv_command@data.csv` output by the
# db2csv command.
#
# Predictions
# ***********
# The :ref:`csv2db_command@Predictions` are in the file
# :ref:`db2csv_command@predict.csv` output by the
# db2csv command.
#
# mtall
# *****
# The omega constraints correspond to
# :ref:`csv2db_command@integrand@mtall` data.
# As a check, we include the mtall data with hold_out equal to one.
#
# Source Code
# ***********
# {xrst_literal
#     BEGIN PYTHON
#     END PYTHON
# }
#
# {xrst_end user_csv2db.py}