Skip to content

tutorial_code_components.md

Latest commit

 

History

History
38 lines (34 loc) · 2.84 KB

tutorial_code_components.md

File metadata and controls

38 lines (34 loc) · 2.84 KB

Components of an EMOD model

EMOD requires a large set of configuration files.

  1. Ingest form: Data/calibration_ingest_form_Nyanza.xlsm
    • Containsx information which the simulation uses as inputs, including
      • List of dynamic parameters which will be fit during the calibration stage
      • List of calibration target variables and how the calibration stage should weight them
      • List of "nodes" or sub-populations in the model, what we use to define geographically isolated metapopulations
      • All calibration target data (empirical data used to calibrate the model)
  2. Input files:
    1. Demographics: InputFiles/Static/Demographics.json
      • Contains births, deaths, and other demographic data
    2. Campaign file: InputFiles/Templates/campaign_Nyanza_baseline_202301.json
      • Defines full structure of the cascade of events which occur in the simulation
      • Together, this set of events contains all events which define
        • How the HIV care cascade occurs
        • How individuals interact with the healthcare system
        • When and how interventions are distributed - testing and treatment
    3. Accessiblity and Risk IP Overlay : InputFiles/Templates/Accessibility_and_Risk_IP_Overlay.json
      • Assigns individual properties to individuals
      • Includes risk levels and accessibility (whether or not person has access to care)
    4. PFA_Overlay: InputFiles/Templates/PFA_Overlay.json
      • Defines pair formation algorithm according to likelihood of pair formation between different demographic groups
    5. Risk_Assortivity_Overlay: InputFiles/Templates/Risk_Assortivity_Overlay.json
      • Defines how different risk groups form partnerships
    6. Configuration file: InputFiles/Templates/config.json
      • Overwrites all other parameter values
      • Useful for formatting outputs, parameterize "reports"
      • (Not all of these parameters are required to build a functioning model, many of them are atavistic cruft relevant to other models that have been copy-pasted through)

Other scripts and files

  1. optim_script.py - this is used to calibrate the simulation, it contains a number of important parameters which are then referenced when the simulation is run
  2. run_scenarios.py - this is a python script used to run scenarios
  3. scenarios.csv - this is a spreadsheet which lists all of the scenarios. You can use this spreadsheet to adjust certain parameters depending on what the scenario needs - used as an input for run_scenarios.py
  4. simtools.ini - this is a shell script which defines the paths to where outputs go, the type of compute nodes that are used to spin up simulation threads
  5. resampled_parameter_sets.csv - spreadsheet of parameter sets, resampled from the distribution of possible parameter values - used as an input for run_scenarios.py