camunda-external-task-client-python3

This repository contains Camunda External Task Client written in Python3.

Implement your BPMN Service Task in Python3.

Python >= 3.7 is required

Installing

Add following line to requirements.txt of your Python project.

git+https://github.com/trustfactors/camunda-external-task-client-python3.git/#egg=camunda-external-task-client-python3

Or use pip to install as shown below:

pip install camunda-external-task-client-python3

Running Camunda with Docker

To run the examples provided in examples folder, you need to have Camunda running locally or somewhere.

To run Camunda locally with Postgres DB as backend, you can use docker-compose.yml file.

$> docker-compose -f docker-compose.yml up

Auth Basic Examples

To run the examples with Auth Basic provided in examples/examplesauth_basic folder, you need to have Camunda with AuthBasic, running locally or somewhere.

To run Camunda with AuthBasic locally with Postgres DB as backend, you can use docker-compose-auth.yml file.

$> docker-compose -f docker-compose-auth.yml up

Usage

• Make sure to have Camunda running.
• Create a simple process model with an External Service Task and define the topic as 'topicName'.
• Deploy the process to the Camunda BPM engine.
• In your Python code:

import time
from camunda.externaltask.externaltask import ExternalTask, TaskResult
from camunda.externaltask.externaltask_worker import ExternalTaskWorker
configuration for the Client
default_config = {
    "maxTasks": 1,
    "lockDuration": 10000,
    "asyncResponseTimeout": 5000,
    "retries": 3,
    "retryTimeout": 5000,
    "sleepSeconds": 30
}
def handle_task(task: ExternalTask) -> TaskResult:     """     This task handler you need to implement with your business logic.     After completion of business logic call either task.complete() or task.failure() or task.bpmn_error()      to report status of task to Camunda     """     # add your business logic here     # ...          # mark task either complete/failure/bpmnError based on outcome of your business logic     failure, bpmnerror = randomtrue(), random_true() # this code simulate random failure     if failure:         # this marks task as failed in Camunda         return task.failure(errormessage="task failed",  errordetails="failed task details",                              maxretries=3, retrytimeout=5000)     elif bpmn_error:         return task.bpmnerror(errorcode="BPMNERRORCODE", error_message="BPMN Error occurred",                                  variables={"var1": "value1", "success": False})          # pass any output variables you may want to send to Camunda as dictionary to complete()     return task.complete({"var1": 1, "var2": "value"})
def random_true():     currentmillitime = int(round(time.time() * 1000))     return currentmillitime % 2 == 0
if name == 'main':    ExternalTaskWorker(workerid="1", config=defaultconfig).subscribe("topicName", handle_task)

About External Tasks

External Tasks are service tasks whose execution differs particularly from the execution of other service tasks (e.g. Human Tasks). The execution works in a way that units of work are polled from the engine before being completed.

camunda-external-task-client-python allows you to create easily such client in Python3.

Features

Start process

Camunda provides functionality to start a process instance for a given process definition.

To start a process instance, we can use startprocess() from engineclient.py

You can find a usage example here.

client = EngineClient()
respjson = client.startprocess(processkey="PARALLELSTEPS_EXAMPLE", variables={"intVar": "1", "strVar": "hello"},
                                 tenantid="6172cdf0-7b32-4460-9da0-ded5107aa977", businesskey=str(uuid.uuid1()))

Fetch and Lock

ExternalTaskWorker(workerid="1").subscribe("topicName", handletask) starts long polling of the Camunda engine for external tasks.

• Polling tasks from the engine works by performing a fetch & lock operation of tasks that have subscriptions. It then calls the handler function passed to subscribe() function. i.e. handle_task in above example.
• Long Polling is done periodically based on the asyncResponseTimeout configuration. Read more about Long Polling.

Complete

from camunda.externaltask.externaltask import ExternalTask, TaskResult
from camunda.externaltask.externaltask_worker import ExternalTaskWorker
def handle_task(task: ExternalTask) -> TaskResult:
    # add your business logic here
    
    # Complete the task
    # pass any output variables you may want to send to Camunda as dictionary to complete()
    return task.complete({"var1": 1, "var2": "value"})
ExternalTaskWorker(workerid="1").subscribe("topicName", handletask)

Handle Failure

from camunda.externaltask.externaltask import ExternalTask, TaskResult
from camunda.externaltask.externaltask_worker import ExternalTaskWorker
def handle_task(task: ExternalTask) -> TaskResult:
    # add your business logic here
    
    # Handle task Failure
    return task.failure(errormessage="task failed",  errordetails="failed task details", 
                        maxretries=3, retrytimeout=5000)
    # This client/worker uses max_retries if no retries are previously set in the task
    # if retries are previously set then it just decrements that count by one before reporting failure to Camunda
    # when retries are zero, Camunda creates an incident which then manually needs to be looked into on Camunda Cockpit
ExternalTaskWorker(workerid="1").subscribe("topicName", handletask)

Handle BPMN Error

from camunda.externaltask.externaltask import ExternalTask, TaskResult
from camunda.externaltask.externaltask_worker import ExternalTaskWorker
def handle_task(task: ExternalTask) -> TaskResult:
    # add your business logic here
    
    # Handle a BPMN Failure
    return task.bpmnerror(errorcode="BPMNERROR", errormessage="BPMN error occurred")
ExternalTaskWorker(workerid="1" ).subscribe("topicName", handletask)

Access Process Variables

from camunda.externaltask.externaltask import ExternalTask, TaskResult
from camunda.externaltask.externaltask_worker import ExternalTaskWorker
def handle_task(task: ExternalTask) -> TaskResult:
    # add your business logic here
    # get the process variable 'score'
    score = task.get_variable("score")
    if int(score) >= 100:
        return task.complete(...)
    else:
        return task.failure(...)
ExternalTaskWorker().subscribe("topicName", handle_task)

Correlate message

Camunda provides functionality to send a message event to a running process instance.

You can read more about the message events here: https://docs.camunda.org/manual/7.13/reference/bpmn20/events/message-events/

In our to send a message event to a process instance, a new function called correlatemessage() is added to engineclient.py

We can correlate the message by:

• processinstanceid
• tenant_id
• business_key
• process_variables

You can find a usage example here.

client = EngineClient()
respjson = client.correlatemessage("CANCELMESSAGE", businesskey="b4a6f392-12ab-11eb-80ef-acde48001122")

AuthBasic Usage

To create an EngineClient with AuthBasic simple

client = EngineClient(config={"auth_basic": {"username": "demo", "password": "demo"}})
respjson = client.startprocess(processkey="PARALLELSTEPS_EXAMPLE", variables={"intVar": "1", "strVar": "hello"},
                                 tenantid="6172cdf0-7b32-4460-9da0-ded5107aa977", businesskey=str(uuid.uuid1()))

To create an ExternalTaskWorker with AuthBasic simple

from camunda.externaltask.externaltask import ExternalTask, TaskResult
from camunda.externaltask.externaltask_worker import ExternalTaskWorker
config = {"auth_basic": {"username": "demo", "password": "demo"}}
def handle_task(task: ExternalTask) -> TaskResult:     # add your business logic here          # Complete the task     # pass any output variables you may want to send to Camunda as dictionary to complete()     return task.complete({"var1": 1, "var2": "value"})
ExternalTaskWorker(workerid="1", config=config).subscribe("topicName", handletask)

License

The source files in this repository are made available under the Apache License Version 2.0.