Python SDK for working with https://github.com/conductor-oss/conductor

Conductor is an open source distributed, scalable and highly available orchestration platform that allows developers to build powerful distributed applications. You can find the documentation for Conductor here: Conductor Docs

Show support for the Conductor OSS. Please help spread the awareness by starring Conductor repo.

• Install SDK
  • Setup SDK
• Start Conductor Server
• Build a conductor workflow based application
  • Step 1: Create a Workflow
  • Step 2: Write Worker
  • Step 3: Write your application
• Implementing Workers
  • Design Principles for Workers
• System Tasks
  • Wait Task
  • HTTP Task
  • Javascript Executor Task
  • Json Processing using JQ
• Executing Workflows
  • Execute workflow asynchronously
  • Execute workflow synchronously
  • Execute dynamic workflows using Code
• Managing Workflow Executions
  • Get the execution status
  • Update workflow state variables
  • Terminate running workflows
  • Retry failed workflows
  • Restart workflows
  • Rerun a workflow from a specific task
  • Pause a running workflow
  • Resume paused workflow
• Searching for workflows
• Handling Failures, Retries and Rate Limits
  • Retries
  • Rate Limits
• Testing your workflows
  • Example Unit Testing application
• Working with Tasks using APIs

Create a virtual environment to build your package

virtualenv conductor
source conductor/bin/activate

SDK needs Python 3.9+.

python3 -m pip install conductor-python

Point the SDK to the Conductor Server API endpoint

export CONDUCTOR_SERVER_URL=http://localhost:8080/api

(Optionally) If you are using a Conductor server that requires authentication

How to obtain the key and secret from the conductor server

export CONDUCTOR_AUTH_KEY=your_key
export CONDUCTOR_AUTH_SECRET=your_key_secret

docker run --init -p 8080:8080 -p 1234:5000 conductoross/conductor-standalone:3.15.0

After starting the server navigate to http://localhost:1234 to ensure the server has started successfully.

Conductor lets you create workflows either in code or using the configuration in JSON that can be created form the code or from the UI. We will explore both the options here.

An application using Conductor uses the following:

1. Workflow: Describes the application's state and how functions are wired. Workflow is what gives your application's code durability and full-blown visualization in the Conductor UI.
2. Worker: Stateless components. Workers can be exposed as HTTP endpoints (aka Microservices) or can be simple task workers implemented using lightweight Conductor SDK in the framework and language of your choice.

Use Code to create workflows

Create greetings_workflow.py with the following:

from conductor.client.workflow.conductor_workflow import ConductorWorkflow
from conductor.client.workflow.executor.workflow_executor import WorkflowExecutor
from examples.greetings import greet

def greetings_workflow(workflow_executor: WorkflowExecutor) -> ConductorWorkflow:
    workflow = ConductorWorkflow(name='hello', executor=workflow_executor)
    workflow >> greet(task_ref_name='greet_ref', name=workflow.input('name'))
    return workflow

(alternatively) Use JSON to create workflows

Create workflow.json with the following:

{
  "name": "hello",
  "description": "hello workflow",
  "version": 1,
  "tasks": [
    {
      "name": "greet",
      "taskReferenceName": "greet_ref",
      "type": "SIMPLE",
      "inputParameters": {
        "name": "${workflow.input.name}"
      }
    }
  ],
  "timeoutPolicy": "TIME_OUT_WF",
  "timeoutSeconds": 60
}

Now, register this workflow with the server:

curl -X POST -H "Content-Type:application/json" \
http://localhost:8080/api/metadata/workflow -d @workflow.json

Create greetings.py with a simple worker and a workflow function.

from conductor.client.worker.worker_task import worker_task


@worker_task(task_definition_name='greet')
def greet(name: str) -> str:
    return f'Hello my friend {name}'

Let's add greetings_main.py with the main method:

from multiprocessing import set_start_method

from conductor.client.automator.task_handler import TaskHandler
from conductor.client.configuration.configuration import Configuration
from conductor.client.http.models import WorkflowRun
from conductor.client.workflow.executor.workflow_executor import WorkflowExecutor
from examples.greetings_workflow import greetings_workflow


def greetings_workflow_run(name: str, workflow_executor: WorkflowExecutor) -> WorkflowRun:
    return workflow_executor.execute(name='hello', version=1, workflow_input={'name': name})


def register_workflow(workflow_executor: WorkflowExecutor):
    workflow = greetings_workflow(workflow_executor=workflow_executor)
    workflow.register(True)

def main():
  
    # points to http://localhost:8080/api by default
    api_config = Configuration()

    workflow_executor = WorkflowExecutor(configuration=api_config)

    # Needs to be done only when registering a workflow one-time
    register_workflow(workflow_executor)

    task_handler = TaskHandler(
        workers=[],
        configuration=api_config,
        scan_for_annotated_workers=True,
        import_modules=['examples.greetings']
    )
    task_handler.start_processes()

    result = greetings_workflow_run('Orkes', workflow_executor)
    print(f'workflow result: {result.output["result"]}')
    task_handler.stop_processes()


if __name__ == '__main__':
    set_start_method('fork')
    main()

The workers can be implemented by writing a simple python function and annotating the function with the @worker_task Conductor workers are services (similar to microservices) that follow Single Responsibility Principle

Workers can be hosted along with the workflow or running a distributed environment where a single workflow uses workers that are deployed and running in different machines/vms/containers. Whether to keep all the workers in the same application or run them as distributed application is a design and architectural choice. Conductor is well suited for both kind of scenarios.

A worker can take inputs which are primitives - str, int, float, bool etc. or can be complex data classes.

Here is an example worker that uses dataclass as part of the worker input.

from conductor.client.worker.worker_task import worker_task
from dataclasses import dataclass

@dataclass
class OrderInfo:
    order_id: int
    sku: str
    quantity: int
    sku_price: float

    
@worker_task(task_definition_name='process_order')
def process_order(order_info: OrderInfo) -> str:
    return f'order: {order_info.order_id}'

Each worker embodies design pattern and follows certain basic principles:

1. Workers are stateless and do not implement a workflow specific logic.
2. Each worker executes a very specific task and produces well-defined output given specific inputs.
3. Workers are meant to be idempotent (or should handle cases where the task that partially executed gets rescheduled due to timeouts etc.)
4. Workers do not implement the logic to handle retries etc, that is taken care by the Conductor server.

System tasks are the pre-built workers that are available in every Conductor server.

System tasks automates the repeated tasks such as calling an HTTP endpoint, executing lightweight ECMA compliant javascript code, publishing to an event broker etc.

Using code to create WAIT task

from conductor.client.workflow.task.wait_task import WaitTask

# waits for 2 seconds before scheduling the next task
wait_for_two_sec = WaitTask(task_ref_name='wait_for_2_sec', wait_for_seconds=2)

# wait until end of jan
wait_till_jan = WaitTask(task_ref_name='wait_till_jsn', wait_until='2024-01-31 00:00 UTC')

# waits until an API call or an event is triggered
wait_for_signal = WaitTask(task_ref_name='wait_till_jan_end')

JSON configuration

{
  "name": "wait",
  "taskReferenceName": "wait_till_jan_end",
  "type": "WAIT",
  "inputParameters": {
    "until": "2024-01-31 00:00 UTC"
  }
}

Make a request to an HTTP(S) endpoint. The task allows making GET, PUT, POST, DELETE, HEAD, PATCH requests.

Using code to create an HTTP task

from conductor.client.workflow.task.http_task import HttpTask

HttpTask(task_ref_name='call_remote_api', http_input={
        'uri': 'https://orkes-api-tester.orkesconductor.com/api'
    })

JSON configuration

{
  "name": "http_task",
  "taskReferenceName": "http_task_ref",
  "type" : "HTTP",
  "uri": "https://orkes-api-tester.orkesconductor.com/api",
  "method": "GET"
}

Execute ECMA compliant Javascript code. Useful when you need to write a script to do data mapping, calculations etc.

from conductor.client.workflow.task.javascript_task import JavascriptTask

say_hello_js = """
function greetings() {
    return {
        "text": "hello " + $.name
    }
}
greetings();
"""

js = JavascriptTask(task_ref_name='hello_script', script=say_hello_js, bindings={'name': '${workflow.input.name}'})

{
  "name": "inline_task",
  "taskReferenceName": "inline_task_ref",
  "type": "INLINE",
  "inputParameters": {
    "expression": " function greetings() {\n  return {\n            \"text\": \"hello \" + $.name\n        }\n    }\n    greetings();",
    "evaluatorType": "graaljs",
    "name": "${workflow.input.name}"
  }
}

jq is like sed for JSON data - you can use it to slice and filter and map and transform structured data with the same ease that sed, awk, grep and friends let you play with text.

from conductor.client.workflow.task.json_jq_task import JsonJQTask

jq_script = """
{ key3: (.key1.value1 + .key2.value2) }
"""

jq = JsonJQTask(task_ref_name='jq_process', script=jq_script)

{
  "name": "json_transform_task",
  "taskReferenceName": "json_transform_task_ref",
  "type": "JSON_JQ_TRANSFORM",
  "inputParameters": {
    "key1": "k1",        
    "key2": "k2",
    "queryExpression": "{ key3: (.key1.value1 + .key2.value2) }",
  }
}

WorkflowClient interface provides all the APIs required to work with workflow executions.

from conductor.client.configuration.configuration import Configuration
from conductor.client.orkes_clients import OrkesClients

api_config = Configuration()
clients = OrkesClients(configuration=api_config)
workflow_client = clients.get_workflow_client() 

Useful when workflows are long-running

from conductor.client.http.models import StartWorkflowRequest

request = StartWorkflowRequest()
request.name = 'hello'
request.version = 1
request.input = {'name': 'Orkes'}
# workflow id is the unique execution id associated with this execution
workflow_id = workflow_client.start_workflow(request)

Useful when workflows complete very quickly - usually under 20-30 second

from conductor.client.http.models import StartWorkflowRequest

request = StartWorkflowRequest()
request.name = 'hello'
request.version = 1
request.input = {'name': 'Orkes'}

workflow_run = workflow_client.execute_workflow(
        start_workflow_request=request, 
        wait_for_seconds=12)

For cases, where the workflows cannot be created statically ahead of the time, Conductor is a powerful dynamic workflow execution platform that lets you create very complex workflows in code and execute them. Useful when the workflow is unique for each execution.

from conductor.client.automator.task_handler import TaskHandler
from conductor.client.configuration.configuration import Configuration
from conductor.client.orkes_clients import OrkesClients
from conductor.client.worker.worker_task import worker_task
from conductor.client.workflow.conductor_workflow import ConductorWorkflow


workflow = ConductorWorkflow(name='dynamic_workflow', version=1, executor=workflow_executor)
get_email = get_user_email(task_ref_name='get_user_email_ref', userid=workflow.input('userid'))
sendmail = send_email(task_ref_name='send_email_ref', email=get_email.output('result'), subject='Hello from Orkes',
                      body='Test Email')
workflow >> get_email >> sendmail

# Execute the workflow and get the workflow run result
result = workflow.execute(workflow_input={'userid': 'usera'})

# Print the workflow status
print(f'workflow completed with status {result.status}')

see dynamic_workflow.py for a fully functional example.

see kitchensink.py for a more complex example.

For more complex workflow example with all the supported features, see kitchensink.py

Workflows represent te application state. With Conductor, you can query the workflow execution state anytime during its lifecycle. You can also send Signals to the workflow that determines the outcome of the workflow state.

WorkflowClient is the client interface used to manage workflow executions.

from conductor.client.configuration.configuration import Configuration
from conductor.client.orkes_clients import OrkesClients

api_config = Configuration()
clients = OrkesClients(configuration=api_config)
workflow_client = clients.get_workflow_client()

The following method lets you query the status of the workflow execution given the id. When the include_tasks is set the response also includes all the completed and in progress tasks.

get_workflow(workflow_id: str, include_tasks: Optional[bool] = True) -> Workflow

Variables inside a workflow are the equivalent to global variables in a program.

update_variables(self, workflow_id: str, variables: dict[str, object] = {})

Terminates a running workflow. Any pending tasks are cancelled and no further work is scheduled for this workflow upon termination. A failure workflow will be triggered, but can be avoided if trigger_failure_workflow is set to False.

terminate_workflow(self, workflow_id: str, reason: Optional[str] = None, trigger_failure_workflow: bool = False)

If the workflow has failed due to one of the task failure after exhausting the retries for the task, the workflow can still be resumed by calling the retry.

retry_workflow(self, workflow_id: str, resume_subworkflow_tasks: Optional[bool] = False)

When a sub-workflow inside a workflow has failed, there are two options:

1. re-trigger the sub-workflow from the start (Default behavior)
2. resume the sub-workflow from the failed task (set resume_subworkflow_tasks to True)

``

A workflow in the terminal state (COMPLETED, TERMINATED, FAILED) can be restarted from the beginning. Useful when retrying from the last failed task is not enough and the whole workflow needs to be started again.

restart_workflow(self, workflow_id: str, use_latest_def: Optional[bool] = False)

In the cases where a worflow needs to be restarted from a specific task rather than from the beginning, re-run provides that option. When issuing the re-run command to the workflow, you have the ability to specify the id of the task from where the workflow should be restarted (as opposed to from the beginning) and optionally, the input of the workflow can also be changed:

rerun_workflow(self, workflow_id: str, rerun_workflow_request: RerunWorkflowRequest)

A running workflow can be put to a PAUSED status. A paused workflow lets the currently running tasks complete, but does not schedule any new tasks until resumed.

pause_workflow(self, workflow_id: str)

Resume operation resumes the currently paused workflow, immediately evaluating its state and scheduling the next set of tasks.

resume_workflow(self, workflow_id: str)

Workflow executions are retained until removed from Conductor. This gives complete visibility into all the executions an application has - regardless of the number of executions. Conductor has a poewrful search API that allows you to search for workflow executions.

search(self, start, size, free_text: str = '*', query: str = None) -> ScrollableSearchResultWorkflowSummary

• free_text: Free text search to look for specific words in the workflow and task input/output
• query SQL like query to search against specific fields in the workflow.

Supported fields for query

Conductor lets you embrace failures rather than worry about failures and complexities that are introduced in the system to handle failures.

All the aspect of handling failures, retries, rate limits etc. are driven by the configuration that can be updated in real-time without having to re-deploy your application.

Each task in Conductor workflow can be configured to handle failures with retries, along with the retry policy (linear, fixed, exponential backoff) and max. number of retry attempts allowed.

See Error Handling for more details.

What happens when a task is operating on a critical resource that can only handle so many requests at a time? Tasks can be configured to have a fixed concurrency (X request at a time) or a rate (Y tasks / time window).

Task Registration

from conductor.client.configuration.configuration import Configuration
from conductor.client.http.models import TaskDef
from conductor.client.orkes_clients import OrkesClients


def main():
    api_config = Configuration()
    clients = OrkesClients(configuration=api_config)
    metadata_client = clients.get_metadata_client()

    task_def = TaskDef()
    task_def.name = 'task_with_retries'
    task_def.retry_count = 3
    task_def.retry_logic = 'LINEAR_BACKOFF'
    task_def.retry_delay_seconds = 1

    # only allow 3 tasks at a time to be in the IN_PROGRESS status
    task_def.concurrent_exec_limit = 3

    # timeout the task if not polled within 60 seconds of scheduling
    task_def.poll_timeout_seconds = 60

    # timeout the task if the task does not COMPLETE in 2 minutes
    task_def.timeout_seconds = 120

    # for the long running tasks, timeout if the task does not get updated in COMPLETED or IN_PROGRESS status in
    # 60 seconds after the last update
    task_def.response_timeout_seconds = 60

    # only allow 100 executions in a 10-second window! -- Note, this is complementary to concurrent_exec_limit
    task_def.rate_limit_per_frequency = 100
    task_def.rate_limit_frequency_in_seconds = 10

    metadata_client.register_task_def(task_def=task_def)

{
  "name": "task_with_retries",
  
  "retryCount": 3,
  "retryLogic": "LINEAR_BACKOFF",
  "retryDelaySeconds": 1,
  "backoffScaleFactor": 1,
  
  "timeoutSeconds": 120,
  "responseTimeoutSeconds": 60,
  "pollTimeoutSeconds": 60,
  "timeoutPolicy": "TIME_OUT_WF",
  
  "concurrentExecLimit": 3,
  
  "rateLimitPerFrequency": 0,
  "rateLimitFrequencyInSeconds": 1
}

Update the task definition:

POST /api/metadata/taskdef -d @task_def.json

See task_configure.py for a detailed working app.

Conductor SDK for python provides a full feature testing framework for your workflow based applications. The framework works well with any testing framework you prefer to use without imposing any specific framework.

Conductor server provide a test endpoint POST /api/workflow/test that allows you to post a workflow along with the test execution data to evaluate the workflow.

The goal of the test framework is as follows:

1. Ability test the various branches of the workflow
2. Confirm the execution of the workflow and tasks given fixed set of inputs and outputs
3. Validate that the workflow completes or fails given specific inputs

Here is example assertions from the test:

...
test_request = WorkflowTestRequest(name=wf.name, version=wf.version,
                                       task_ref_to_mock_output=task_ref_to_mock_output,
                                       workflow_def=wf.to_workflow_def())
run = workflow_client.test_workflow(test_request=test_request)

print(f'completed the test run')
print(f'status: {run.status}')
self.assertEqual(run.status, 'COMPLETED')

...

See test_workflows.py for a fully functional example on how to test a moderately complex workflow with branches.