Programmer’s Guide

In This Section

• Tutorial
• Key Concepts
• Monitoring Jobs
• Working With Task Results
• Troubleshooting
• Internals

Simple Example

Here is a very simple example application. It adds two numbers together and writes the result to the console.

from agiparallel.client import ClusterJobScheduler
from agiparallel.constants import CommonResources, JobPriority, Operator


class SimpleTask():
    def __init__(self, first, second):
        self.a = first
        self.b = second

    def execute(self):
        self.result = self.a + self.b


class MyEnvironment():
    def setup(self):
        pass

    def teardown(self):
        pass


def simple_task_example():
    with ClusterJobScheduler("localhost") as scheduler:
        scheduler.connect()
        job = scheduler.create_job()

        # add one or more tasks
        job.add_task(SimpleTask(1, 1))

        # set the task environment
        job.task_environment = MyEnvironment()

        # set various job options
        job.name = "My job name"
        job.description = "My job description"
        job.max_interrupted_retry_attempts = 0
        job.priority = JobPriority.LOW
        job.add_task_preconditions(CommonResources.AVAILABLE_CORES, Operator.GREATER_THAN, 2)

        # submit task to the job scheduler
        job.submit()

        # wait for results
        job.wait_until_done()

        print("The result is: ", job.tasks[0].result)


if __name__ == "__main__":
    simple_task_example()

Copy and paste the code into a python file and run. The application first connects and submits a task to the coordinator. Next, an agent will be assigned the work and execute the task in an isolated process. Finally, the result of the task is sent back to the client.

If you’re using an IDE like JetBrains PyCharm, step through the code line by line. Examine the messages written from the coordinator and agent consoles. Take note of how the coordinator and agent respond after stepping through each line. For example, stepping past job.submit() will cause the coordinator console to log “added 1 tasks for job #1 to queue”. What this means will be discussed later.

Continue to experiment with this application. Try to add more tasks to the job, set the jobs’s task_environment, or modify some of the job options. As this example shows, the basic structure and core pieces of an application written with the API are simple but powerful!

Next Steps

With a rudimentary example application finished, explore the help system and move on to more advanced topics. The following sections give various topics in the API more treatment.

Title	Read this if you want to…
Tutorial	Read a step-by-step narrative of another sample application.
How To	Get solutions to common patterns in the API.
Key Concepts	Understand the key concepts and building blocks of the product.
Frequently Asked Questions	Get answers to frequently asked questions.
Internals	Gain a firm understanding of the internals of the product.
Library Reference	Explore what is available in the API.