workflows-emulator 1.0.1

Google Workflows Emulator and testing utilities

Google Cloud Workflows Emulator

This is both a library and cli tool.

Using the CLI

Calling a workflow

1. Create a workflow config for your project

   # small_config.workflow.yaml
   main:
     params: [ table_name ]
     steps:
       - assign_variables:
           assign:
             - table_parts: ${text.split(table_name, ".")}
             - project_id: ${sys.get_env("GOOGLE_CLOUD_PROJECT_ID")}
             - dataset_id: ${table_parts[-2]}
             - table_id: ${table_parts[-1]}
             - config:
                 project_id: ${project_id}
                 dataset_id: ${dataset_id}
                 table_id: ${table_id}
             - query: |
                 SELECT column_name
                 FROM `${project_id}.${dataset_id}.INFORMATION_SCHEMA.COLUMNS`
                 WHERE table_name = '${table_id}'
       - final:
           return: ${query}
   

   ---

   NOTE 🟢 Remember to use the Cloud Workflows Json schema in your IDE to get the correct syntax highlighting, autocompletion and error checking.

2. Define your environment variables in the .env file. Alternatively you can pass a custom .env file to the emulator with the --env-file flag.

   # .env
   GOOGLE_CLOUD_PROJECT_ID=numbers-crunching-123
   

3. To run a single workflow in the CLI:

   workflows-emulator \
     --config test/data/small_config.workflow.yaml \
     run \
     --data='{"var_content": "lowercase text"}'
   

   To start the server emulating the Google Cloud service:

   workflows-emulator --config test/data/small_config.workflow.yaml serve
   

   and then call the server with a POST request:

   curl --request 'POST' \
      --header 'Content-Type: application/json' \
      --data '{"argument": "{\"var_content\": \"hello\"}"}' \
      'http://localhost:8000/v1/projects/my_project/locations/europe-west4/workflows/small_config/executions'
   

4. The output will be printed to the console

   Log step HELLO
   "HELLO"
   

Using the library for testing

import os
from workflows_emulator.main import load_workflow, execute_workflow, execute_subworkflow, get_step, execute_step

def test_execute_main_workflow():
    os.environ['GOOGLE_CLOUD_PROJECT_ID'] = 'my_project'
    config = load_workflow('path/to/workflow.yaml')
    params = {'my_var': 3}
    result = execute_workflow(config, params)
    assert result == 5


def test_execute_sub_workflow():
    os.environ['GOOGLE_CLOUD_PROJECT_ID'] = 'my_project'
    config = load_workflow('path/to/workflow.yaml')
    params = {'sub_workflow_param': 3}
    result = execute_subworkflow(config, params)
    assert result == 5

def test_step():
    os.environ['GOOGLE_CLOUD_PROJECT_ID'] = 'my_project'
    config = load_workflow('path/to/workflow.yaml')
    main_workflow = config['main']
    context = {'sub_workflow_param': 3}
    ok, _index, step_config  = get_step(main_workflow['steps'], context)
    assert ok
    _context, _next, result = execute_step(step_config, context)
    assert result == 5

---

Reason behind this emulator

How to develop and debug your workflows according to Google Cloud

Running a Workflow goes like this:

WORKFLOW_NAME=my-workflow-name
WORKFLOW_FILE_PATH=my_workflow.yaml

gcloud workflows deploy ${WORKFLOW_NAME} \
  --location=europe-west4 \
  --call-log-level=log-all-calls \
  --source=${WORKFLOW_FILE_PATH}

gcloud workflows run ${WORKFLOW_NAME} \
  --location=europe-west4 \
  --call-log-level=log-all-calls

Update connectors

Workflows provides a set of connectors to interact with Google Cloud REST APIs easier. They are listed in the documentation. If new connectors are added, they can be refreshed running:

get-discovery-documents

Then open a PR with the newly generated files.

Not implemented (yet)

Some of the std lib modules are not implemented as the behavior is difficult to mimic, or it is work-in-progress:

• experimental.executions — it's use is discouraged in the docs
• events — callbacks are complex to run locally