Create dispatch_computed_file_generation command

[avrohomgottlieb]

Context

As mentioned in Issue #840, a dispatch_computed_file_generation command must be build, which will be responsible for dispatching file computation jobs to Batch.

Problem or idea

The dispatch_computed_file_generation command should query the Project model for all projects in the database that do not have any ComputedFiles associated with them, iterate over those projects and then submit them to the Batch queue one by one.

The query could look as follows:

Project.objects.filter(project_computed_files__is_null=True)

Solution or next step

from django.core.management.base import BaseCommand

import boto3

from scpca_portal.config.logging import get_and_configure_logger

batch = boto3.client('batch', region_name=os.environ["AWS_REGION"])
logger = get_and_configure_logger(__name__)


class Command(BaseCommand):
    def handle(self, *args, **kwargs) -> None:
        self.dispatch_computed_file_generation()
  
    def submit_batch_job(self, project_id:str) -> None:
        response = batch.submit_job(
            jobName='job-name',
            jobQueue='job-queue',
            jobDefinition='job-definition',
            containerOverrides={
                'command': ['python', 'manage.py', 'generate_computed_files', 'project.scpca_id']            
         }
         logger.info(f'{project} submitted to Batch as job {response["jobId"]}')
        )

    def dispatch_computed_file_generation(self) -> None:
        for project in Project.objects.filter(project_computed_files__is_null=True):
            self.submit_batch_job(project.scpca_id)

From an error handling perspective, submit_batch_job could return a bool (indicating whether or not the job was successfully added to the queue). If the job was not successfully added to the job queue, one tactic could be to try submitting the job again. A possible way of doing that could be as follows:

from queue import Queue

dispatch_queue = Queue()
for project in Project.objects.filter(project_computed_files__is_null=True):
    dispatch_queue.put(project)    

while not dispatch_queue.empty():
    project = dispatch_queue.get()
    if not self.submit_batch_job(project.scpca_id):
        dispatch_queue.put(project)

[avrohomgottlieb]

Updates for this Issue:

• This command should take an optional project argument. If this argument is not passed, then it should submit_jobs for computed_file generation for all projects that don't have computed_files associated with them.
• submit_job should set command to python manage.py generate_computed_file --project/sample [scpca_id] --config_name [config name]

[avrohomgottlieb]

Addressed in PR #936