Workflow

Running JUNE_NZ involves three main steps:

Step 1: Creating the input dataset.
Step 2: Running the model.
Step 3: Generating the model diagnosis.

Each step utilizes the output generated by the previous step as input.

Note that the JUNE_NZ environment must be enabled when running the worklfow (e.g., conda activate june_nz), and we may also need to specify the PYTHONPATH accordingly (e.g., export PYTHONPATH=/home/zhangs/Github/June_NZ).

Step 1. Creating the input dataset

The wrapper for creating the input dataset is cli/cli_data.py, which can be triggered as:

python cli_data.py --workdir <WORKING DIRECTORY>
                   --cfg <DATA CREATION CONFIGURATION>
                   [--scale <POPULATION PERCENTAGE>]
                   [--disease_cfg_dir <DISEASE CONFIGURATION DIRECTORY>]
                   [--policy_cfg_path <POLICY CONFIGURATION FILE>]
                   [--vaccine_cfg_path <VACCINATION CAMPAIGN CONFIGURATION FILE>]
                   [--simulation_cfg_path <SIMULATION CONFIGURATION FILE>]
                   [--use_sa3_as_super_area]

where:

--workdir: Working directory.

--cfg: Data configuration.

--scale: Population percentage to be used in the model (e.g., 0.1 means that we only use 10% of NZ population).

--disease_cfg_dir: Disease configuration directory, which must contain population comorbidities, infection probability, infection outcome, symptom trajectory and virus intensity.

--policy_cfg_path: Policy configuration file.

--vaccine_cfg_path: Vaccine compaign configuration file.

--simulation_cfg_path: Simulation configuration file.

--use_sa3_as_super_area: If using SA3 as super area, otherwise use the regional councils.

For example, the input data can be created by python cli_data.py --workdir etc/data/realworld_auckland --cfg etc/cfg/run/june_data.yml --scale 0.1 --disease_cfg_dir etc/cfg/disease/covid-19 --policy_cfg_path etc/cfg/policy/policy1.yaml --vaccine_cfg_path etc/cfg/disease/vaccine/vaccine1.yaml --simulation_cfg_path etc/cfg/simulation/simulation_cfg.yml --use_sa3_as_super_area.

Step 2. Running the model

The model can be run using cli/cli_june.py, which can be triggered as:

python cli_june.py --workdir <WORKING DIRECTORY>
                   --cfg <MODEL CONFIGURATION>
                   [--tuning_cfg <TUNING CONFIGURATION>]

where:

--workdir: Working directory.

--cfg: Model running configuration.

--tuning_cfg: Model tuning configuration [Default: None]. If not set then no tuning will be done from the input data.

For example, the model diagnosis can be created by: python cli_june.py --workdir /tmp/june_realworld_auckland_base --cfg etc/cfg/run/june_nz2.yml.

Step 3. Generating the model diagnosis

The wrapper for creating the model output diagnosis is cli/cli_diags.py, which can be triggered as:

python cli_diags.py --workdir <WORKING DIRECTORY>
                    --cfg <JUNE MODEL DIAGNOSIS CONFIGURATION PATH>
                    --june_data_dir <JUNE MODEL OUTPUT DIRECTORY>

where:

--workdir: Working directory.

--cfg: Model diagnosis configuration.

--june_data_dir: June model output directory

For example, the model diagnosis can be created by python cli_diags.py --workdir etc/data/june_realworld_auckland_base_diag --cfg etc/cfg/run/june_diags.yml --june_data_dir /tmp/june_realworld_auckland_base/output.