django-openacct 0.0.5

Installation

django-openacct can be installed from PyPI, and requires no external dependencies aside from Django.

?> pip install django-openacct

Quick Start

The following shows a small setup script being piped into the Django shell interpreter:

?> cat openacct_example.py

from openacct.models    import *
from openacct.shortcuts import *

# Create a System and Service for a cluster
cluster = System.objects.create(name='cluster', description='A Cluster')
service = Service.objects.create(
    name='cluster-core-hours', units='core-hours', system=cluster,
    charge_rate=0.05, description='CPU core usage * wallclock hours'
)

# Create Users
bob = User.objects.create(name='bob')
sue = User.objects.create(name='sue')
tux = User.objects.create(name='tux')

# Create Projects and grant access to the cluster's service
phys = create_project('phys', pi=bob, description='Physics Dept')
chem = create_project('chem', pi=sue, description='Chemsitry Dept')
grant_service_access(service, project=phys)
grant_service_access(service, project=chem)

# Add a user to the projects
add_user_to_project(tux, chem)
add_user_to_project('tux', 'phys')


?> ./manage.py shell < openacct_example.py

This example shows a basic setup for a center, with a single compute resource which is charged on a per core-hour basis. Several users and two projects, one for a Physics Department, and another for a Chemistry Department, are created.

Access to use a service is managed through Accounts, which are being implicitly created for each project. When passing the project parameter to grant_service_access, all active accounts are granted the same access. Accounts can be created explicitly and granted access individually when desired.

Finally, the user tux is added to both of the created projects, demonstrating how the shortcut functions will accept both the name of an object or the object itself for all parameters referencing one of the models in the schema.

With this setup in place, job records could be added to the database as Job objects, with Transactions making use of the cluster-core-hour service linked to them.