csv2db_command#

View page source

Conversion of a Csv File to a Dismod_at Database#

Deprecated#

This command was deprecated on 2021-12-05.

Syntax#

As Program#

dismodat.py database csv2db configure_csv measure_csv

As Python Function#

dismod_at.csv2db_command ( database , configure_csv , measure_csv )

See Also#

get_started

Example#

This command has limited capability and is only meant as an example to help one get started using dismod_at. See user_csv2db.py for an example that uses this command.

mtall#

The all cause mortality data mtall in the measure_csv file has special meaning. We assume it has been converted to constraints on other cause mortality; see mtother below. For this reason, it should not be included when fitting, and is only in measure_csv as a check that the desired values are satisfied (by checking residuals in a data fit). For this reason, hold_out must be one for all the mtall data; i.e., it is not included during a fit.

mtother#

Constraint#

The other cause mortality data mtother in the measure_csv file has special meaning. It is intended to represent the all cause mortality data as constraints on other cause mortality in the model. To be specific, the meas_value is a constraint on \(\omega\) at the corresponding age-time pairs.

Age-Time Pairs#

Each row with integrand = mtother must have age_lower == age_upper and time_lower == time_upper .

Rectangular Grid#

The mtother data must be specified on a rectangular grid; i.e., each age that appears, appears in one and only one row for every time that appears. This property can also be stated as each time that appears, appears in one and only one row for every age that appears.

hold_out#

The hold_out must be one for all the mtother data because it is a constraint, not data, during a fit.

Rate Grid#

All of the non zero rates use the age-time grid corresponding the mtother data. In other words, they are modeled as piecewise bilinear between the age-time points at which mtother is specified.

Predictions#

The avgint_table is set up so that predictions for the integrands Sincidence , remission , mtexcess , corresponding to the value of the rates iota, rho, chi on the Rectangular Grid . Only the non-zero rates are included. Predictions for prevalence in the rectangular grid are also included.

database#

This argument is an str containing the name of the dismod_at database file that is written by this command.

configure_csv#

This argument is an str containing the configuration file name and must end with the .csv extension. The first row contains the following column names name , value . Column names that begin with c_ are comments and will not be used by future versions of csv2db . The configuration options are documented by the corresponding name below.

non_zero_rates#

The value in this row is a list rates that are non-zero in the model. The possible rates are pini , iota , rho , chi , omega . The rates in the list are separated by a single space and omega must appear in the list. There is no default value for this value; i.e., it must appear.

measure_csv#

is an str containing the data file name and must end with the .csv extension. Each row of the data file corresponds to one data point. The first row of the file contains the column names. The other rows correspond to data points. Each column of the necessary columns in the data file is documented under its column name below. Column names that begin with c_ are comments and will not be used by future versions of csv2db .

integrand#

This column of measure_csv contains one of the valid integrands:

Sincidence#

The incidence rate relative to susceptible population: \(\iota\).

remission#

The remission rate: \(\rho\).

mtexcess#

The excess mortality rate: \(\chi\).

mtother#

The other cause mortality rate: \(\omega\).

mtwith#

The with condition mortality rate: \(\omega + \chi\).

susceptible#

The susceptible fraction of the population: \(S\).

withC#

The with condition fraction of the population: \(C\).

prevalence#

The prevalence of the condition: \(P = C / [ S + C ]\).

Tincidence#

The incidence rate relative to the total population: \(\iota [ 1 - P ]\).

mtspecific#

The cause specific mortality rate: \(\chi P\).

mtall#

The all cause mortality rate: \(\omega + \chi P\).

mtstandard#

The standardized mortality ratio: \([ \omega + \chi ] / [ \omega + \chi P ]\).

relrisk#

The relative risk: \([ \omega + \chi ] / \omega\).

age_lower#

This column of measure_csv contains The initial age for averaging the integrand for this row; \(b\).

age_upper#

This column of measure_csv contains the final age for averaging the integrand for this row; \(c\).

time_lower#

This column of measure_csv contains the initial time for averaging the integrand for this row; \(r\).

time_upper#

This column of measure_csv contains the final time for averaging the integrand for this row; \(s\).

meas_value#

This column of measure_csv contains the value of the average integrand plus measurement noise

\[e + \frac{1}{c-b} \frac{1}{s-r} \int_b^c \int_r^s I(a , t) \; da \; dt\]

where \(e\) is the measurement noise, \(a\) and \(t\) are the age and time integration variables, and \(I(a, t)\) is the value of the integrand for this row.

meas_std#

This column of measure_csv contains the standard deviation of the measurement noise \(e\).

hold_out#

This column of measure_csv contains either zero or one. If it is one, this row is included dismod_at fits. Otherwise it is excluded. In either case, the residuals are computed for this row.