lilota 0.0.5

A lightweight solution for long-running tasks

lilota

Table of contents

• What is it?
• When to use it?
• How to install it?
• How to use it?

What is it?

lilota is a lightweight solution for long-running tasks. When we talk about such a solution we're essentially speaking about a tool that can handle task management efficiently without the overhead and complexity of other systems like Celery, RabbitMQ, etc.

When to use it?

Lets assume you have a web application and you send a request to the server where you want to run long running tasks like processing images, generating reports based on large data sets, sending emails etc.

In such cases you do not want to let the user wait for the response of the server. Instead you want to send the request to the server and let the user know that the request has been received and that the server is working on it.

Here lilota can help you by processing the long-running task in a separate process and returning the response to the user as soon as the task has been started.

How to install it?

Binary installers for the latest released version are available at the Python Package Index (PyPI).

pip install lilota

How to use it?

Implement the task you want to execute

First we create a class where we implement our task logic. This class must derive from TaskBase and must contain a run method.

Example:

from lilota.models import TaskBase

...

class LongRunningTask(TaskBase):

  def run(self):
    self.logger.info("Wait 10 seconds to simulate a long running task")
    time.sleep(10)

    self.set_output(self.task_info.id, {
      "result": self.task_info.input["number1"] + self.task_info.input["number2"]
    })

    self.logger.info("Task finished")

In this example we create the task LongRunningTask. In the run method we wait 10 seconds and then we calculate the sum of two numbers (number1 and number2). For the moment it is not important to understand where number1 and number2 are coming from. We come to that very soon. The result is in the end stored in a dictionary and here the key result is used. You are free to use whatever name you want for that key and you can use how many keys and values you want of course.

What you can see is that we also log messages here. lilota is using the build-in logging module and this is already integrated inside the base class TaskBase and can be directly used.

Create a store

lilota needs a store where all the information about the running tasks are stored. Several store implementations are of course possible. By default lilota comes with a store that holds its data inside the memory. In the future we will also create a store that connects to a database and is using Django.

Here is an example for our in-memory store:

from lilota.stores import MemoryTaskStore, StoreManager

...

# Create store
StoreManager.register("Store", MemoryTaskStore)
store_manager = StoreManager()
store_manager.start()
store = store_manager.Store()

StoreManager derives from the BaseManager that comes from the multiprocessing module in Python. The reason why this is used is to run the store inside its own process in order to access it from all the started processes. More information about such Managers can be found here.

After the manager is started we get the store from it (store = store_manager.Store()). With that store we can now create a TaskRunner instance.

Create a TaskRunner instance

The heart of lilota is the TaskRunner class. It can be instantiated like this:

from lilota.runner import TaskRunner

...

runner = TaskRunner(store)

The only mandatory parameter is the store object we created before.

Register tasks and start the task runner

As soon as we have our TaskRunner object we can register our tasks we want to execute. This registration process does not mean that we run that task but we just make lilota aware of our task we want to execute.

runner.register("long_running_task", LongRunningTask)
runner.start()

Two parameters are needed here:

• name: The name of our task. That name is used later to execute a task.
• class: Here we just add our class that we created above.

As soon as all tasks are registered we can start the task runner. This also does not mean we start all the registered tasks. It only means we start the task runner and the task runner is waiting for the tasks.

Lets feed the task runner

Now that we have a running instance of our TaskRunner, we want to execute our created task. In order to do that you have to execute the following:

runner.add("long_running_task", "This task adds two numbers", {"number1": 1, "number2": 2})

Here we add a task to the runner by specifying the name ("long_running_task") that we used before when we registered the task. The second argument is a description we can add. And the third argument is a dictionary with the input parameters. Here in this example we pass a dictionary with the two keys number1 and number2 and their values. These input parameters can be used then in the task itself. You can put whatever you want in that dictionary.

Data inside the store

As soon as the task is running information about the task are stored in the store. We can get for example all tasks with the following command:

tasks = store.get_all_tasks()
print(f"Result: { tasks[0].output["result"] }")

This is a very simplyfied version that would fail when the output is not already set. Remember lilota is a task runner that executes tasks in different processes and the store is available everywhere. So it is possible that the task is saved in the store but still running. Below we print a complete example where we wait until the task is done.

Stop lilota

When the application is closed and lilota is not needed anymore you can stop it and close the store.

runner.stop()
store_manager.shutdown()

You have to do it in that order. When lilota gets stopped it is waiting for tasks that are still running.

The complete example:

from lilota.runner import TaskRunner
from lilota.models import TaskBase
from lilota.stores import MemoryTaskStore, StoreManager
import logging
import time


class LongRunningTask(TaskBase):

  def run(self):
    self.logger.info("Start a long running task")
    self.logger.info("Wait 10 seconds to simulate a long running task")
    time.sleep(1)
    self.logger.info("Task finished")

    # Set result
    self.set_output(self.task_info.id, {
      "result": self.task_info.input["number1"] + self.task_info.input["number2"]
    })
    

if __name__ == "__main__":
  # Create store
  StoreManager.register("Store", MemoryTaskStore)
  store_manager = StoreManager()
  store_manager.start()
  store = store_manager.Store()

  # Create task runner
  runner = TaskRunner(store)

  # Register task
  runner.register("long_running_task", LongRunningTask)

  # Start the task runner
  runner.start()

  # Run a task
  runner.add("long_running_task", "This task adds two numbers", {"number1": 1, "number2": 2})

  # Get result from the store (1 + 2 = 3)
  tasks = None
  while True:
    tasks = store.get_all_tasks()
    if tasks[0].progress_percentage >= 100:
      break

  if hasattr(tasks[0], 'output'):
    print(f"Result: { tasks[0].output["result"] }")

  # Stop the task runner
  runner.stop()

  # Close the store because it is running in a separate process
  store_manager.shutdown()