models.py

from django.db import models
from django.urls import reverse
from django.utils.timezone import datetime
from django.db.models import Sum

# constants
datetime_format_string = '%Y-%m-%dT%H:%M:%SZ'

class Workflow(models.Model):
    workflow_uri = models.CharField(unique=True, max_length=30, blank=True, null=True)
    repository = models.CharField(max_length=100, blank=True, null=True)
    commit_id = models.CharField(max_length=15, blank=True, null=True)
    path = models.CharField(max_length=100, blank=True, null=True)
    oi_size_fraction = models.FloatField(blank=True, null=True)
    meta_scheduling = models.JSONField(null=True, blank=True)

    def __str__(self):
        return str(self.id)


class Task(models.Model):

    # Task control properties
    task_type = models.CharField(max_length=20, default="regular")
    filter = models.CharField(max_length=30, blank=True, null=True)
    environment = models.JSONField(null=True, blank=True)
    new_status = models.CharField(max_length=50, default="defining", null=True)
    status = models.CharField(db_index=True, default="unknown", max_length=50,blank=True, null=True)

    resume = models.BooleanField(verbose_name="Resume", default=True)
    creationTime = models.DateTimeField(verbose_name="CreationTime",default=datetime.utcnow, blank=True)

    priority = models.IntegerField(default=100, null=True)
    purge_policy = models.CharField(max_length=5, default="no", blank=True, null=True)
    stage_request_id = models.IntegerField(null=True)

    # LOFAR properties
    project = models.CharField(max_length=100, blank=True, null=True, default="unknown")
    sas_id = models.CharField(verbose_name="SAS_ID",max_length=15, blank=True, null=True)
    inputs = models.JSONField(null=True, blank=True)
    outputs = models.JSONField(null=True, blank=True)
    metrics = models.JSONField(null=True, blank=True)

    meta_scheduling = models.JSONField(null=True, blank=True)

    size_to_process = models.PositiveBigIntegerField(default=0, null=True, blank=True)
    size_processed = models.PositiveBigIntegerField(default=0, null=True, blank=True)
    total_processing_time = models.IntegerField(default=0, null=True, blank=True)


    # relationships
    workflow = models.ForeignKey(Workflow, related_name='tasks', on_delete=models.SET_NULL, null=True, blank=True)
    predecessor = models.ForeignKey('self', related_name='successors', on_delete=models.SET_NULL, null=True, blank=True)

    def __str__(self):

        return str(self.id) + ' - (' + self.task_type + ') - ' + str(self.sas_id)

    # this translates a view-name (from urls.py) back to a url, to avoid hardcoded url's in the html templates
    # bad : <td><a href="/atdb/taaks/{{ task.id }}/" target="_blank">{{ task.taskID }} </a> </td>
    # good: <td><a href="{{ task.get_absolute_url }}" target="_blank">{{ task.taskID }} </a> </td>
    def get_absolute_url(self):
        return reverse('task-detail-view-api', kwargs={'pk': self.pk})

    @property
    def predecessor_status(self):
        try:
            return self.predecessor.status
        except:
            return "no_predecessor"

class LogEntry(models.Model):
    cpu_cycles = models.IntegerField(null=True,blank=True)
    wall_clock_time = models.IntegerField(null=True,blank=True)
    url_to_log_file = models.CharField(max_length=100, blank=True, null=True)
    service = models.CharField(max_length=30, blank=True, null=True)
    step_name = models.CharField(max_length=30, blank=True, null=True)
    timestamp = models.DateTimeField(blank=True, null=True)
    status = models.CharField(max_length=50,default="defined", blank=True, null=True)
    description = models.CharField(max_length=100, blank=True, null=True)
    size_processed = models.PositiveBigIntegerField(default=0, null=True, blank=True)
    # relationships
    task = models.ForeignKey(Task, related_name='log_entries', on_delete=models.CASCADE, null=False)

    def __str__(self):
        return str(self.id)+ ' - '+ str(self.task)+' - '+self.status


class Status(models.Model):
    name = models.CharField(max_length=50, default="unknown")
    timestamp = models.DateTimeField(default=datetime.utcnow, blank=True)

    # relationships
    task = models.ForeignKey(Task, related_name='status_history', on_delete=models.CASCADE, null=False)

    # the representation of the value in the REST API
    def __str__(self):
        formatedDate = self.timestamp.strftime(datetime_format_string)
        return str(self.name)+' ('+str(formatedDate)+')'


class Configuration(models.Model):
    filter = models.CharField(max_length=30, blank=True, null=True)
    key = models.CharField(max_length=50)
    value = models.CharField(max_length=255)

    # the representation of the value in the REST API
    def __str__(self):
        return str(self.key)


class Job(models.Model):
    type = models.CharField(db_index=True, max_length=20, default=None,null=True, blank=True)
    task_id = models.IntegerField(null=True, blank=True)
    job_id = models.IntegerField(null=True, blank=True)
    metadata = models.JSONField(null=True, blank=True)

    # the representation of the value in the REST API
    def __str__(self):
        return 'task_id:'+str(self.task_id)+', job_id:'+str(self.job_id)


class PostProcessingRule(models.Model):
    aggregation_key = models.CharField(db_index=True, max_length=20, default=None,null=True, blank=True)
    trigger_status = models.CharField(db_index=True, default="unknown", max_length=50, blank=True, null=True)

    # relationships
    workflow_to_process = models.ForeignKey(Workflow, related_name='to_process', on_delete=models.SET_NULL, null=True, blank=True)
    workflow_to_apply = models.ForeignKey(Workflow, related_name='to_apply',on_delete=models.SET_NULL, null=True, blank=True)

    # the representation of the value in the REST API
    def __str__(self):
        return str(self.aggregation_key)+' - '+str(self.trigger_status)


class LatestMonitor(models.Model):
    name = models.CharField(max_length=50, default="unknown")
    type = models.CharField(max_length=20, default="ldv-service", null=True, blank=True)
    timestamp = models.DateTimeField(default=datetime.utcnow, blank=True)
    hostname = models.CharField(max_length=50, default="unknown")
    process_id = models.IntegerField(null=True, blank=True)
    description = models.CharField(max_length=255, blank=True, null=True)
    status = models.CharField(max_length=50, default="ok", null=True)
    metadata = models.JSONField(null=True, blank=True)

    # the representation of the value in the REST API
    def __str__(self):
        return str(self.name) + ' - ('+ self.hostname+') - '+str(self.timestamp) + ' - ' + self.status


class Monitor(models.Model):
    name = models.CharField(max_length=50, default="unknown")
    type = models.CharField(max_length=20, default="ldv-service", null=True, blank=True)
    timestamp = models.DateTimeField(default=datetime.utcnow, blank=True)
    hostname = models.CharField(max_length=50, default="unknown")
    process_id = models.IntegerField(null=True, blank=True)
    description = models.CharField(max_length=255, blank=True, null=True)
    status = models.CharField(max_length=50, default="ok", null=True)
    metadata = models.JSONField(null=True, blank=True)

    def save(self, *args, **kwargs):
        # check if this combination of service name + hostname already exists
        # in the LatestMonitor, and update if it is newer.
        try:
            latestMonitor = LatestMonitor.objects.get(name=self.name,hostname=self.hostname)
            latestMonitor.delete()
        except:
            pass

        # this combination of name and hostname didn't yet exist, create it.
        latestMonitor = LatestMonitor(
            name=self.name,
            type=self.type,
            timestamp=self.timestamp,
            hostname = self.hostname,
            process_id = self.process_id,
            description = self.description,
            status = self.status,
            metadata = self.metadata
        )
        latestMonitor.save()

        # finally save the Monitor info itself also
        super(Monitor, self).save(*args, **kwargs)

    # the representation of the value in the REST API
    def __str__(self):
        return str(self.name) + ' - ('+ self.hostname+') - '+str(self.timestamp) + ' - ' + self.status