m-v-nikolov / dtk-tools-py3

Documentation available: http://institutefordiseasemodeling.github.io/dtk-tools

The dtk package is intended to strengthen and simplify the interaction between researchers and the EMOD model.

Modules contained in this package are intended to:

• Empower the user to configure diverse simulations and arbitrarily complex campaigns built up from a standardized library of configuration fragments and utility functions;
• Facilitate transparent switching between local and remote HPC commissioning, job-status queries, and output analysis;
• Enable the configuration of arbitrary collections of simulations (e.g. parameteric sweeps) through an extensible library of builder classes;
• Collect a library of post-processing analysis functionality, e.g. filtering, mapping, averaging, plotting.

Installation

To install the dtk-tools, first clone the repository:

git clone https://github.com/InstituteforDiseaseModeling/dtk-tools.git

Make sure you have Python 3.6 x64 installed (available here).

From a command-prompt, run the following from the dtk-tools directory:

python setup.py

Note: If pip command is not found on your system, make sure to add the Python scripts directory (by default in Windows: C:\Python27\Scripts) to your PATH environment variable.

To test if dtk-tools is correctly installed on your machine issue a:

dtk -h

If the command succeed and present you with the details of the dtk command you are all set!

MAC users

Please refer to MacOSX install instructions for more information.

CentOS7 users

Please refer to CentOS install instructions for more information.

Setup

To configure your user-specific paths and settings for local and HPC job submission, please create a simtools.ini file in the same folder that contains your scripts or modify the master simtools.ini at dtk-tools/simtools/simtools.ini

Simulation job management is handled through the various dtk command-line options, e.g. dtk run example_sweep or dtk analyze example_plots. For a full list of options, execute dtk --help. Many example configurations for simulation sweeps and analysis processing may be found in the examples directory.