This is a sandbox that allows you to play around with Temporal using a Go workflow worker and a Python activity worker. It implements a trip booking saga (read more about the Saga pattern) similar to the one described in Temporal Booking Saga Tutorial in PHP. It showcases how you might use workflow and activity workers written in different languages.

• ./docker-compose.yaml runs Temporal server, including our workflow and activity workers. It exposes Temporal Web UI at localhost:8088. It also runs a temporal-development container that has Go and Python dependencies installed and code mounted. You can use this container to run Temporal CLI tool tctl commands, run workflows and activities, modify code, run tests. If you are using VS Code, you might find it useful to connect to this container with Remote Containers extensions.

• ./Dockerfile - temporal-development image definition.

• ./workflow-worker - Go implementation of a workflow worker implementing a trip booking saga.

• ./activity-worker - Python implementation of activities that the booking saga workflow uses. This also contains workflow_tests.py - these are examples of live tests that run the booking workflow and provide different activity implementations to show how the workflow acts.

• ./run-tests.sh - this is a script that runs live tests. It creates a test Temporal namespace, runs a test workflow worker and runs Python workflow_tests.py.

Running this example depends only on Docker and docker-compose. Run the containers:

docker-compose up -d

Then connect to the development container:

docker exec -it temporal-development /bin/bash

Check Temporal cluster health:

tctl cluster health

The output should look something like this:

temporal.api.workflowservice.v1.WorkflowService: SERVING

You can run the booking workflow with Temporal CLI tool, tctl:

tctl workflow run \
    --taskqueue booking-workflows \
    --workflow_type book-trip \
    --workflow_id YOUR-WORKFLOW-ID \
    --input '"YOUR-USER-ID"'

--workflow_id is optional. If you omit it, the ID will be generated randomly. '"YOUR-USER-ID"' is double-quoted because input values must be JSON values.

You can inspect the workflow events at http://localhost:8088/namespaces/default/workflows.

There are two kinds of tests in this repository: Go unit tests and Python live tests.

Go unit tests are at ./workflow-worker/trip_workflow_test.go. They are light, they use Go Temporal SDK's testsuite and mock the activities. To run these tests:

cd /temporal/workflow-worker
go test

Lear more about Go Temporal testing in Temporal Go SDK testing docs.

Python tests (./activity-workflow/workflow_tests.py) are more involved. They run the book-trip workflow on a real Temporal server but provide different activity implementations for each test and inspect the result. These tests require:

• A running Temporal server which has a namespace test created.
• A workflow worker running in that namespace.

You can create the test namespace, run the workflow worker and run the tests just by running a helper script:

/temporal/run-tests.sh

Alternatively, you can do this by hand:

1. tctl --namespace test namespace register
2. cd /temporal/workflow-worker && go run . --namespace test --host temporal:7233 & (running as a background task)
3. cd /temporal/activity-worker && pytest workflow_tests.py

After the tests are run, you can inspect the workflows at http://localhost:8088/namespaces/test/workflows. You should see one workflow that succeeded and one forkflow that failed with a message Failed to complete bookings. All bookings canceled..

If you want to use the examples without the Docker, you'll need:

• A running Temporal server (how-to).
• Go installed (how-to).
• Python 3.10 and poetry installed (how-to).

To run the workflow worker:

cd ./workflow-worker
go build
./worker --namespace <default> --host <host:port>

To run the activity worker:

cd ./activity-worker
poetry install
poetry run python main.py --namespace <default> --host <host:port>