python

Argo Python SDK

This is the Python SDK for Argo Workflows.

Warning

This SDK is deprecated and will be removed in version 3.7 of Argo Workflows. We recommend migrating to Hera.

Requirements

Python >= 3.6

Installation

If you'd like to install the official releases of the SDK on PyPI, please run the following:

pip install argo-workflows

Otherwise, you can install the latest development version of the SDK via the following:

pip install git+https://github.com/argoproj/argo-workflows@main#subdirectory=sdks/python/client

If you have any questions regarding which specific SDK version to use, please see the section on versioning.

Getting Started

You can submit a workflow from a raw YAML like the following:

from pprint import pprint

import requests
import yaml

import argo_workflows
from argo_workflows.api import workflow_service_api
from argo_workflows.model.io_argoproj_workflow_v1alpha1_workflow_create_request import (
    IoArgoprojWorkflowV1alpha1WorkflowCreateRequest,
)

configuration = argo_workflows.Configuration(host="https://127.0.0.1:2746")
configuration.verify_ssl = False

resp = requests.get('https://raw.githubusercontent.com/argoproj/argo-workflows/main/examples/hello-world.yaml')
manifest = yaml.safe_load(resp.text)

api_client = argo_workflows.ApiClient(configuration)
api_instance = workflow_service_api.WorkflowServiceApi(api_client)
api_response = api_instance.create_workflow(
    namespace="argo",
    body=IoArgoprojWorkflowV1alpha1WorkflowCreateRequest(workflow=manifest, _check_type=False),
    _check_return_type=False)
pprint(api_response)

Note that _check_type=False is required here to avoid type checks against manifest which is a Python dictionary and _check_return_type=False avoids an existing issue with OpenAPI generator.

Alternative, you can submit a workflow with an instance of IoArgoprojWorkflowV1alpha1Workflow constructed via the SDK like the following:

from pprint import pprint

import argo_workflows
from argo_workflows.api import workflow_service_api
from argo_workflows.model.container import Container
from argo_workflows.model.io_argoproj_workflow_v1alpha1_template import IoArgoprojWorkflowV1alpha1Template
from argo_workflows.model.io_argoproj_workflow_v1alpha1_workflow import IoArgoprojWorkflowV1alpha1Workflow
from argo_workflows.model.io_argoproj_workflow_v1alpha1_workflow_create_request import (
    IoArgoprojWorkflowV1alpha1WorkflowCreateRequest,
)
from argo_workflows.model.io_argoproj_workflow_v1alpha1_workflow_spec import (
    IoArgoprojWorkflowV1alpha1WorkflowSpec,
)
from argo_workflows.model.object_meta import ObjectMeta

configuration = argo_workflows.Configuration(host="https://127.0.0.1:2746")
configuration.verify_ssl = False

manifest = IoArgoprojWorkflowV1alpha1Workflow(
    metadata=ObjectMeta(generate_name='hello-world-'),
    spec=IoArgoprojWorkflowV1alpha1WorkflowSpec(
        entrypoint='whalesay',
        templates=[
            IoArgoprojWorkflowV1alpha1Template(
                name='whalesay',
                container=Container(
                    image='docker/whalesay:latest', command=['cowsay'], args=['hello world']))]))

api_client = argo_workflows.ApiClient(configuration)
api_instance = workflow_service_api.WorkflowServiceApi(api_client)

if __name__ == '__main__':
    api_response = api_instance.create_workflow(
        namespace='argo',
        body=IoArgoprojWorkflowV1alpha1WorkflowCreateRequest(workflow=manifest),
        _check_return_type=False)
    pprint(api_response)

Examples

You can find additional examples here.

API Reference

You can find the API reference here.

Versioning

When the Python SDK was migrated to this repository, we kept the argo-workflows name for the Python package since we wanted to publish it to the same package on PyPI. However, the existing argo-workflows package was on version 5.0.0 already whereas Argo Workflows was still on 3.x.x.

In order to make it easier indicate backwards compatibility between the SDK version and the Argo Workflows releases, we bump the Python SDK major version by 3 and keep the minor and patch versions. For example, Python SDK with version 6.3.0rc2 is released with Argo Workflows v3.3.0-rc2. Note that the hyphen in -rc2 is removed intentionally to follow Python package versioning conventions documented in PEP-0440.