Expose root workflow execution #805

THardy98 · 2025-03-28T01:55:27Z

What was changed

Expose root workflow execution to workflow info (available from workflow info) and workflow execution info (available from workflow describe/list)

Closes [Feature Request] Expose root workflow execution #774
How was this tested:
Updating existing info,describe,list tests.
Any docs updates needed?
Probably

temporalio/api/batch/v1/__init__.py

temporalio/client.py

cretz · 2025-03-28T13:34:05Z

temporalio/client.py

+    root_id: Optional[str]
+    """ID for the root workflow"""
+
+    root_run_id: Optional[str]
+    """Run ID for the root workflow"""


So technically this is an incompatible change since users can technically instantiate this object themselves which some users may do in a mock client situation (I know they do in other languages). Two options:

Add a note to the docstring of this class (and WorkflowExecutionDescription) that the constructor may change in backwards incompatible ways

Add defaults for these values

I'm on the fence on which I prefer. @dandavison or @THardy98 - preference?

I think I'd lean toward the latter while we can and introduce the former when we have to. I suppose default values for these fields would be None?

Added defaults

Removed defaults - linter complains about having fields with default values before fields without defaults, and this class is inherited by WorkflowExecutionDescription below, making it annoying to fix. In any case, does it not default to None already, being Optional ?

Just move the fields to the bottom. I know this breaks our alphabetical order, but new fields with defaults being at the bottom makes the most sense anyways, we just happen to have alphabetical order because we have never added a field with a default before. No, that the data type is Optional is unrelated to whether the field has to be specified on construction or not.

Yes but it still conflicts with the non-defaults fields in WorkflowExecutionDescription, unless I'm misunderstanding you

Oh, I see, even moving to the bottom of WorkflowExecution doesn't help with. Yeah if we were on Python > 3.10, I'd say mark the dataclasses as kw_only=True but since we can't do that, let's (for now) just leave them where they are and go with option 1 from my comment at the top of this thread which is "Add a note to the docstring of this class (and WorkflowExecutionDescription) that the constructor may change in backwards incompatible ways".

cretz · 2025-03-28T13:35:12Z

temporalio/workflow.py

@@ -427,6 +427,7 @@ class Info:
    headers: Mapping[str, temporalio.api.common.v1.Payload]
    namespace: str
    parent: Optional[ParentInfo]
+    root: Optional[RootInfo]


Can we have a test that confirms this is populated (and that it is populated on result of handle.describe() from client side)?

Added a test in test_workflow and added an assertion to a describe test in test_client

… release

THardy98 · 2025-03-28T22:59:39Z

skipping the test in time-skipping environments, needs temporalio/sdk-java#2441
which I suppose hasn't been released yet (or is not used in the Python SDK yet)

cretz · 2025-03-31T14:36:38Z

tests/helpers/__init__.py

@@ -150,3 +156,25 @@ async def admitted_update_task(
        lambda: workflow_update_exists(client, handle.id, id),
    )
    return update_task
+
+
+async def child_started(


This does not seem related to child. Maybe a assert_workflow_exists_eventually helper instead? Also, this does not need to accept the workflow (just use get_workflow_handle which is untyped), but when changing to assert_workflow_exists_eventually you may want this so you can return a handle for the workflow instead of a bool.

changed to assert_workflow_exists_eventually returns a WorkflowHandle (if exists)

cretz · 2025-03-31T14:42:33Z

tests/worker/test_workflow.py

+@workflow.defn(name="ChildWorkflowInfo")
+class ChildWorkflowInfo:
+    def __init__(self) -> None:
+        self.blocked = True
+
+    @workflow.signal
+    def unblock(self):
+        self.blocked = False
+
+    @workflow.run
+    async def run(self):
+        await workflow.wait_condition(lambda: not self.blocked)
+        return workflow.info().root
+
+
+@workflow.defn(name="WithChildWorkflowInfo")
+class WithChildWorkflowInfo:


Suggested change

@workflow.defn(name="ChildWorkflowInfo")

class ChildWorkflowInfo:

def __init__(self) -> None:

self.blocked = True

@workflow.signal

def unblock(self):

self.blocked = False

@workflow.run

async def run(self):

await workflow.wait_condition(lambda: not self.blocked)

return workflow.info().root

@workflow.defn(name="WithChildWorkflowInfo")

class WithChildWorkflowInfo:

@workflow.defn

class ExposeRootChildWorkflow:

def __init__(self) -> None:

self.blocked = True

@workflow.signal

def unblock(self):

self.blocked = False

@workflow.run

async def run(self):

await workflow.wait_condition(lambda: not self.blocked)

return workflow.info().root

@workflow.defn

class ExposeRootWorkflow:

Don't need to manually set name on these, and would rather name the workflow similar to its test if at all possible

removed and renamed

cretz · 2025-03-31T14:43:50Z

tests/worker/test_workflow.py

+        self.blocked = False
+
+    @workflow.run
+    async def run(self):


Set return type hints on all non-test methods

cretz · 2025-03-31T14:44:38Z

tests/worker/test_workflow.py

+            "Java test server needs release with: https://github.com/temporalio/sdk-java/pull/2441"
+        )
+    async with new_worker(client, WithChildWorkflowInfo, ChildWorkflowInfo) as worker:
+        child_wf_id = "child-wf-id"


Taking this ID globally is not very safe ID for multiple executions of this test environment. Can you add a UUID in here somewhere?

cretz

Looks great, nothing blocking (still may be worth that dataclass constructor warning to users about constructor changing)

THardy98 · 2025-03-31T19:54:38Z

TYFR :)

THardy98 requested a review from a team as a code owner March 28, 2025 01:55

cretz reviewed Mar 28, 2025

View reviewed changes

THardy98 force-pushed the expose_root_wf_exec branch from bf61320 to 36eea79 Compare March 28, 2025 21:37

exposing root info to workflow info and workflow execution info

fe9aac1

THardy98 force-pushed the expose_root_wf_exec branch from 36eea79 to c94eb3b Compare March 28, 2025 21:48

added test

ab70dfa

THardy98 force-pushed the expose_root_wf_exec branch from c94eb3b to ab70dfa Compare March 28, 2025 22:03

skip test in time-skipping env, added link for PR that needs to be in…

951f2ce

… release

cretz reviewed Mar 31, 2025

View reviewed changes

improve test helper, changes to test wf name, return types, and ids

93aa061

THardy98 requested a review from cretz March 31, 2025 18:31

cretz approved these changes Mar 31, 2025

View reviewed changes

THardy98 merged commit 92b7758 into main Mar 31, 2025
14 checks passed

THardy98 deleted the expose_root_wf_exec branch March 31, 2025 19:54

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Expose root workflow execution #805

Expose root workflow execution #805

THardy98 commented Mar 28, 2025

cretz Mar 28, 2025

THardy98 Mar 28, 2025

cretz Mar 28, 2025

THardy98 Mar 28, 2025

THardy98 Mar 28, 2025 •

edited

Loading

cretz Mar 31, 2025 •

edited

Loading

THardy98 Mar 31, 2025

cretz Mar 31, 2025

cretz Mar 28, 2025

THardy98 Mar 28, 2025

THardy98 commented Mar 28, 2025

cretz Mar 31, 2025

THardy98 Mar 31, 2025

cretz Mar 31, 2025

THardy98 Mar 31, 2025

cretz Mar 31, 2025

THardy98 Mar 31, 2025

cretz Mar 31, 2025

THardy98 Mar 31, 2025

cretz left a comment

THardy98 commented Mar 31, 2025

Expose root workflow execution #805

Expose root workflow execution #805

Conversation

THardy98 commented Mar 28, 2025

What was changed

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

THardy98 Mar 28, 2025 • edited Loading

Choose a reason for hiding this comment

cretz Mar 31, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

THardy98 commented Mar 28, 2025

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

cretz left a comment

Choose a reason for hiding this comment

THardy98 commented Mar 31, 2025

THardy98 Mar 28, 2025 •

edited

Loading

cretz Mar 31, 2025 •

edited

Loading