Tasks - doing heavy work in the background

Introduction

Background processing using threads is a way to maintain a responsive user interface when heavy processing is going on. Tasks can be used to achieve threading in QGIS.

A task (QgsTask) is a container for the code to be performed in the background, and the task manager (QgsTaskManager) is used to control the running of the tasks. These classes simplify background processing in QGIS by providing mechanisms for signaling, progress reporting and access to the status for background processes. Tasks can be grouped using subtasks. See the QGIS Python API doc for details on methods, signals and attributes.

The global task manager (found with QgsApplication.taskManager()) is normally used. This means that your tasks may not be the only tasks that are controlled by the task manager.

There are several ways to create a QGIS task:

  • Create your own task by extending QgsTask

    class SpecialisedTask(QgsTask):
    
  • Create a task from a function

    QgsTask.fromFunction(u'heavy function', heavyFunction,
                         onfinished=workdone)
    
  • Create a task from a processing algorithm

    QgsProcessingAlgRunnerTask(u'native:buffer', params, context,
                               feedback)
    

Warning

Any background task (regardless of how it is created) must NEVER perform any GUI based operations, such as creating new widgets or interacting with existing widgets. Qt widgets must only be accessed or modified from the main thread. Attempting to use them from background threads will result in crashes.

Dependencies between tasks can be described using the addSubTask function of QgsTask. When a dependency is stated, the task manager will automatically determine how these dependencies will be executed. Wherever possible dependencies will be executed in parallel in order to satisfy them as quickly as possible. If a task on which another task depends is canceled, the dependent task will also be canceled. Circular dependencies can make deadlocks possible, so be careful.

If a task depends on a layer being available, this can be stated using the setDependentLayers function of QgsTask. If a layer on which a task depends is not available, the task will be canceled.

Once the task has been created it can be scheduled for running using the addTask function of the task manager. Adding a task to the manager automatically transfers ownership of that task to the manager, and the manager will cleanup and delete tasks after they have executed. The scheduling of the tasks is influenced by the task priority, which is set in addTask.

The status of tasks can be monitored using QgsTask and QgsTaskManager signals and functions.

Examples

Extending QgsTask

In this example RandomIntegerSumTask extends QgsTask and will generate 100 random integers between 0 and 500 during a specified period of time. If the random number is 42, the task is aborted and an exception is raised. Several instances of RandomIntegerSumTask (with subtasks) are generated and added to the task manager, demonstrating two types of dependencies.

import random
from time import sleep

from qgis.core import (
    QgsApplication, QgsTask, QgsMessageLog,
    )

MESSAGE_CATEGORY = 'RandomIntegerSumTask'

class RandomIntegerSumTask(QgsTask):
    """This shows how to subclass QgsTask"""
    def __init__(self, description, duration):
        super().__init__(description, QgsTask.CanCancel)
        self.duration = duration
        self.total = 0
        self.iterations = 0
        self.exception = None
    def run(self):
        """Here you implement your heavy lifting.
        Should periodically test for isCanceled() to gracefully
        abort.
        This method MUST return True or False.
        Raising exceptions will crash QGIS, so we handle them
        internally and raise them in self.finished
        """
        QgsMessageLog.logMessage('Started task "{}"'.format(
                                     self.description()),
                                 MESSAGE_CATEGORY, Qgis.Info)
        wait_time = self.duration / 100
        for i in range(100):
            sleep(wait_time)
            # use setProgress to report progress
            self.setProgress(i)
            arandominteger = random.randint(0, 500)
            self.total += arandominteger
            self.iterations += 1
            # check isCanceled() to handle cancellation
            if self.isCanceled():
                return False
            # simulate exceptions to show how to abort task
            if arandominteger == 42:
                # DO NOT raise Exception('bad value!')
                # this would crash QGIS
                self.exception = Exception('bad value!')
                return False
        return True
    def finished(self, result):
        """
        This function is automatically called when the task has
        completed (successfully or not).
        You implement finished() to do whatever follow-up stuff
        should happen after the task is complete.
        finished is always called from the main thread, so it's safe
        to do GUI operations and raise Python exceptions here.
        result is the return value from self.run.
        """
        if result:
            QgsMessageLog.logMessage(
                'Task "{name}" completed\n' \
                'Total: {total} (with {iterations} '\
              'iterations)'.format(
                  name=self.description(),
                  total=self.total,
                  iterations=self.iterations),
              MESSAGE_CATEGORY, Qgis.Success)
        else:
            if self.exception is None:
                QgsMessageLog.logMessage(
                    'Task "{name}" not successful but without '\
                    'exception (probably the task was manually '\
                    'canceled by the user)'.format(
                        name=self.description()),
                    MESSAGE_CATEGORY, Qgis.Warning)
            else:
                QgsMessageLog.logMessage(
                    'Task "{name}" Exception: {exception}'.format(
                        name=self.description(),
                        exception=self.exception),
                    MESSAGE_CATEGORY, Qgis.Critical)
                raise self.exception
    def cancel(self):
        QgsMessageLog.logMessage(
            'Task "{name}" was canceled'.format(
                name=self.description()),
            MESSAGE_CATEGORY, Qgis.Info)
        super().cancel()


longtask = RandomIntegerSumTask('waste cpu long', 20)
shorttask = RandomIntegerSumTask('waste cpu short', 10)
minitask = RandomIntegerSumTask('waste cpu mini', 5)
shortsubtask = RandomIntegerSumTask('waste cpu subtask short', 5)
longsubtask = RandomIntegerSumTask('waste cpu subtask long', 10)
shortestsubtask = RandomIntegerSumTask('waste cpu subtask shortest', 4)

# Add a subtask (shortsubtask) to shorttask that must run after
# minitask and longtask has finished
shorttask.addSubTask(shortsubtask, [minitask, longtask])
# Add a subtask (longsubtask) to longtask that must be run
# before the parent task
longtask.addSubTask(longsubtask, [], QgsTask.ParentDependsOnSubTask)
# Add a subtask (shortestsubtask) to longtask
longtask.addSubTask(shortestsubtask)

QgsApplication.taskManager().addTask(longtask)
QgsApplication.taskManager().addTask(shorttask)
QgsApplication.taskManager().addTask(minitask)

Task from function

Create a task from a function (run in this example). The first parameter of the function will hold the QgsTask for the function. An important (named) parameter is on_finished, that specifies a function that will work on the result. The run function in this example has an additional named parameter wait_time.

import random
from time import sleep

MESSAGE_CATEGORY = 'TaskFromFunction'

def doSomething(task, wait_time):
    """
    Raises an exception to abort the task.
    Returns a result if success.
    The result will be passed together with the exception (None in
    the case of success) to the on_finished method
    """
    QgsMessageLog.logMessage('Started task {}'.format(task.description()),
                             MESSAGE_CATEGORY, Qgis.Info)
    wait_time = wait_time / 100
    total = 0
    iterations = 0
    for i in range(100):
        sleep(wait_time)
        # use task.setProgress to report progress
        task.setProgress(i)
        arandominteger = random.randint(0, 500)
        total += arandominteger
        iterations += 1
        # check task.isCanceled() to handle cancellation
        if task.isCanceled():
            stopped(task)
            return None
        # raise an exception to abort the task
        if arandominteger == 42:
            raise Exception('bad value!')
    return {'total': total, 'iterations': iterations,
            'task': task.description()}

def stopped(task):
    QgsMessageLog.logMessage(
        'Task "{name}" was canceled'.format(
            name=task.description()),
        MESSAGE_CATEGORY, Qgis.Info)

def completed(exception, result=None):
    """This is called when doSomething is finished.
    Exception is not None if doSomething raises an exception.
    Result is the return value of doSomething."""
    if exception is None:
        if result is None:
            QgsMessageLog.logMessage(
                'Completed with no exception and no result '\
                '(probably manually canceled by the user)',
                MESSAGE_CATEGORY, Qgis.Warning)
        else:
            QgsMessageLog.logMessage(
                'Task {name} completed\n'
                'Total: {total} ( with {iterations} '
                'iterations)'.format(
                    name=result['task'],
                    total=result['total'],
                    iterations=result['iterations']),
                MESSAGE_CATEGORY, Qgis.Info)
    else:
        QgsMessageLog.logMessage("Exception: {}".format(exception),
                                 MESSAGE_CATEGORY, Qgis.Critical)
        raise exception

# Creae a few tasks
task1 = QgsTask.fromFunction(u'Waste cpu 1', doSomething,
                             on_finished=completed, wait_time=4)
task2 = QgsTask.fromFunction(u'Waste cpu 2', dosomething,
                             on_finished=completed, wait_time=3)
QgsApplication.taskManager().addTask(task1)
QgsApplication.taskManager().addTask(task2)

Task from a processing algorithm

Create a task that uses the algorithm qgis:randompointsinextent to generate 50000 random points inside a specified extent. The result is added to the project in a safe way.

from functools import partial
from qgis.core import (QgsTaskManager, QgsMessageLog,
                       QgsProcessingAlgRunnerTask, QgsApplication,
                       QgsProcessingContext, QgsProcessingFeedback,
                       QgsProject)

MESSAGE_CATEGORY = 'AlgRunnerTask'

def task_finished(context, successful, results):
    if not successful:
        QgsMessageLog.logMessage('Task finished unsucessfully',
                                 MESSAGE_CATEGORY, Qgis.Warning)
    output_layer = context.getMapLayer(results['OUTPUT'])
    # because getMapLayer doesn't transfer ownership, the layer will
    # be deleted when context goes out of scope and you'll get a
    # crash.
    # takeMapLayer transfers ownership so it's then safe to add it
    # to the project and give the project ownership.
    if output_layer and output_layer.isValid():
        QgsProject.instance().addMapLayer(
             context.takeResultLayer(output_layer.id()))

alg = QgsApplication.processingRegistry().algorithmById(
                                      u'qgis:randompointsinextent')
context = QgsProcessingContext()
feedback = QgsProcessingFeedback()
params = {
    'EXTENT': '0.0,10.0,40,50 [EPSG:4326]',
    'MIN_DISTANCE': 0.0,
    'POINTS_NUMBER': 50000,
    'TARGET_CRS': 'EPSG:4326',
    'OUTPUT': 'memory:My random points'
}
task = QgsProcessingAlgRunnerTask(alg, params, context, feedback)
task.executed.connect(partial(task_finished, context))
QgsApplication.taskManager().addTask(task)

See also: https://www.opengis.ch/2018/06/22/threads-in-pyqgis3/.