PandemicExerciseSimulator

This is a Python command line implementation of a pandemic exercise simulator using a SEATIRD compartment model.

Install:

$ git clone https://github.com/TACC/PandemicExerciseSimulator
$ pip install .

Test:

$ make run
$ make debug
$ pytest

Input Data Required:

The simulator requires the data from the following folder:

.
└── data
    └── texas
        ├── INPUT.json
        ├── contact_matrix.5
        ├── county_age_matrix.5
        ├── high_risk_ratios.5
        ├── nu_value_matrix.5
        ├── relative_susceptibility.5
        ├── vaccine_adherence.5
        ├── vaccine_effectiveness.5
        └── work_matrix_rel.csv

• INPUT.json: Simulation properties file (see schema)
• contact_matrix.5: 5x5 matrix of contact ratios between age groups
• county_age_matrix.5: Populations for each county divided into age groups
• high_risk_ratios.5: List of risk ratio for each age group
• nu_value_matrix.5: Nx4 columns (N=num of age groups) low/high death rate x low/high risk. Nu is the transmitting (asymptomatic/treatable/infectious) to deceased rate
• relative_susceptibility.5 List of relative susceptibility for each age group
• vaccine_adherence.5: List of vaccine adherences for each age group
• vaccine_effectiveness.5: List of vaccine effectiveness for each age group
• work_matrix_rel.csv: NxN matrix (N=num of counties) for travel flow

Parameters Required:

Among other things, the following parameters are expected to be defined in the INPUT.json file:

• R0: (ex: 1.8) Reproduction number
• beta_scale: (ex: 65) R0 correction factor - R0 is divided by this value and stored as beta
• tau: (ex: 0.83333333) exposed to asymptomatic rate
• kappa: (ex: 0.52631579) asymptomatic to treatable rate
• gamma: (ex: 0.24390244) transmitting (asymptomatic/treatable/infectious) to recovered rate
• chi: (ex: 1.0) treatable to infectious rate
• nu_high: ("yes" or "no") use high or low death rates
• vaccine_wastage_factor: (ex: 60) half the vaccine stockpile will be wasted every N days
• antiviral_effectiveness: (ex: 0.8) antiviral effectiveness factor
• antiviral_wastage_factor: (ex: 60) half the antiviral stockpile will be wasted every N days

Notes

• contact_matrix.5, county_age_matrix.5, and work_matrix_rel.csv: all taken from original data
• vaccine_effectiveness.5: used to be in properties file under 'params', but the five age groups were hardcoded. Pulled this out into its own file
• vaccine_adherence.5: used to be in properties file under 'params', but the five age groups were hardcoded. Pulled this out into its own file
• relative_susceptibility.5: was hardcoded as SIGMA in ModelParameters.cpp, but pulled it out and made it its own file.
• high_risk_ratios.5: was hardcoded in ModelParameters.cpp, but pulled it out and made it its own file
• relative_susceptibility.5: was hardcoded as nu[][] in ModelParameters.cpp, took this out and made it its own file
• Params (chi=1.0): Chi was originally hardcoded in ModelParameters.cpp, took this out and made it a param.
• Params (vaccine_wastage_factor=60): was orginally hardcoded in ModelParameters.cpp, took this out and made it a param. "Every N days half the stock pile is wasted"
• Params (antiviral_wastage_factor=60): was orginally hardcoded in ModelParameters.cpp, took this out and made it a param. "Every N days half the stock pile is wasted"
• Params (antiviral_effectievness=0.8): was orginally hardcoded in ModelParameters.cpp, took this out and made it a param.

Future Development Notes

• Needs functionality to scroll to a certain date in time, change parameters, then continue run from there.
• Model should be checkpointable and show provenance of how it arrived there
• Should be able to compare counties easily
• Implement proper schema for inputs and outputs https://github.com/python-jsonschema/jsonschema