Basic Usage¶
autojob has three main functionalities:
coordinator - Codeless workflow design
advance & next - Semi-automatic workflow management
harvest - Data aggregation
coordinator¶
coordinator is a codeless interface for computational chemistry workflow design,
management, and parallelization. With coordinator, you can design workflows,
parametrize tasks, use custom templates, and dump task inputs to dedicated
directories.
advance & next¶
advance & next are two subpackages dedicated to semi-automatic
workflow management. These subpackages can be used to step forward in
a workflow from the command-line (with advance.advance) or
restart a calculation after modifying inputs (with next.relaxation).
harvest¶
The harvest.harvest subpackage provides utilities for collecting the
results of completed tasks into structured data model that is suitable
for database storage. All harvested results are collected in
task.Task objects which can be dumped into JSON-able dictionaries
with autojob.task.Task.to_directory().