Running

Overview

AFO requires Python and the required packages to be installed on your device (see the Getting Started section) and can be run using any IDLE or through the teminal. The model is run from RunAfoRaw_Multi.py but the trials which are run and the application of any sensitivities are controlled using exp.xlsx (see below). To customise the execution you can pass in optional args to RunAfoRaw_Multi.py <experiment number> <processors> which control the experiment to be run and the number of processors used. If no arguments are passed experiment type 0 will be executed using 1 processor. Once the model has been run two output files per trial are saved to the pkl folder.

Running the reports is much like running the model (note to generate a report the trial must have been previously run), simply just run RunReportsRaw.py rather than RunAfoRaw_Multi.py. As with running the core model arguments <experiment number> <processors> <report number> can be passed in. If no arguments are passed experiment type 0 will be reported using 1 processor and the default report settings. For further information on reporting see the report section.

IDLE

Scripts can be run by any python interpreter. Vscode is a popular example however Pycharm preforms better for debugging large scrips such as the stock module.

Run using cmd/terminal

The model can be run through the cmd/terminal. This is used for cloud computing. To run the script on terminal; navigate to the directory with the model and run the script using the following command: python RunAfoRaw_Multi.py <experiment number> <processors> The python used is the one listed in PATH but you can use others by specifying the path when running the command eg /PATH/python RunAfoRaw_Multi.py <experiment number> <processors>. A similar process can be done using the anaconda prompt.

Exp.xlsx

Firstly some Terminology.

Analysis:

An analysis is all the lines of Exp.xlsx and may also include the same Exp.xlsx being used on multiple regional models.

Experiment:

An experiment is a block of Exp.xlsx that is examining an individual question pertaining to that analysis

Trial:

A trial is a single line within an experiment.

Exp.xlsx is used to control the sensitivity values for each trial. It also controls if the trial is run, the level of output and if the trial is reported. Trials are grouped into experiments in the ‘Exp Group’ column. Trials with the same number are in the same experiment. The user can then select which experiment to run by passing in a number when running the scripts from terminal (e.g. python RunAfoRaw_Multi.py 2 this will run all the trials with a 2 in the ‘Exp Group’ column). If no argument is passed in then all trials are included. In Pycharm you can pass in an argument by right clicking on the module name, then there is an option to ‘Create’ (it changes to Edit once yo have used it). In there is a parameter box which you put a value in. Within an experiment you can use True/False in the ‘Run Trials’ column to control if a given trial is executed.