Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat: implement async client for LROs #707

Merged
merged 14 commits into from
Oct 7, 2024

Conversation

parthea
Copy link
Collaborator

@parthea parthea commented Sep 24, 2024

Towards b/362946071

This PR depends on #700 and #701

@product-auto-label product-auto-label bot added the size: xl Pull request size is extra large. label Sep 24, 2024
@parthea parthea changed the base branch from main to operations-rest-async-transport September 24, 2024 00:24
Copy link

conventional-commit-lint-gcf bot commented Sep 24, 2024

🤖 I detect that the PR title and the commit message differ and there's only one commit. To use the PR title for the commit history, you can use Github's automerge feature with squashing, or use automerge label. Good luck human!

-- conventional-commit-lint bot
https://conventionalcommits.org/

@parthea parthea force-pushed the operations-rest-async-client branch 3 times, most recently from 1492985 to cb2eab0 Compare September 25, 2024 15:32
@parthea parthea changed the base branch from operations-rest-async-transport to main September 25, 2024 17:35
@parthea parthea force-pushed the operations-rest-async-client branch from 787f0ba to 35d88d7 Compare September 25, 2024 17:55
@parthea parthea changed the title [WIP] Operations rest async client feat: Add AbstractOperationsAsyncClient Sep 25, 2024
@ohmayr ohmayr force-pushed the operations-rest-async-client branch from 5ea5e27 to 4d75cca Compare October 1, 2024 20:30
@ohmayr ohmayr changed the base branch from main to operations-rest-async-transport October 1, 2024 20:31
@ohmayr ohmayr force-pushed the operations-rest-async-client branch 2 times, most recently from 5e6c365 to 74e9bb8 Compare October 1, 2024 21:02
@ohmayr ohmayr changed the base branch from operations-rest-async-transport to main October 1, 2024 21:03
@ohmayr ohmayr marked this pull request as ready for review October 1, 2024 21:09
@ohmayr ohmayr requested review from a team as code owners October 1, 2024 21:09
@ohmayr ohmayr force-pushed the operations-rest-async-client branch from 4f3671c to 2af34aa Compare October 2, 2024 20:45
@ohmayr ohmayr changed the base branch from main to operations-rest-async-transport October 2, 2024 21:26
Base automatically changed from operations-rest-async-transport to async-rest-lro-support-in-core October 3, 2024 05:06
@ohmayr ohmayr changed the title feat: Add AbstractOperationsAsyncClient feat: implement async client for LROs Oct 3, 2024
@ohmayr ohmayr force-pushed the operations-rest-async-client branch from 5564d01 to f448498 Compare October 3, 2024 05:19
Copy link
Contributor

@vchudnov-g vchudnov-g left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I have some initial comments. I'll continue next week.

The main thing I want to work out in my head is the class hierarchy, and how it pertains to {sync, async} × {grpc, rest}.

What I've worked out so far is this class hierarchy, but I don't see where gRPC fits in. I need to think about this more, and it might be worth a discussion:


# do any of these include gRPC?

AbstractOperationsBaseClient   o----  OperationsRestTransport
   ^ ^                         o----  OperationsRestAsyncTransport
   | |       
   | |       
   | AbstractOperationsClient       # sync
   |
   |
   AbstractOperationsAsyncClient

ListOperationsPagerBase
  ^  ^
  |  |
  |  |
  |  ListOperationsPager           # sync
  |
  |
  ListOperationsAsyncPager

google/api_core/operations_v1/__init__.py Outdated Show resolved Hide resolved
google/api_core/operations_v1/__init__.py Outdated Show resolved Hide resolved
super().__init__(
credentials=credentials, # type: ignore
# NOTE: If a transport is not provided, we force the client to use the async
# REST transport, as it should.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

  • Why should it? This abstract operations class does not say it's REST-specific, so why couldn't we go with gRPC async? We should see default to one of the ones that are supported (have the right dependencies); if both gRPC and REST are possible, sure, we could have a policy as to which one we choose for the default. As per my other comment, if the plan is to do this as we clean up this code, etc., let's add a TODO with link to a tracking issue that includes setting the right default.

  • Regardless, I would remove the final clause:

Suggested change
# REST transport, as it should.
# REST transport.
  • Also, update the class doc above so instead if "If set to None, a transport is chosen automatically", it says "If set to None, this defaults to 'rest_asyncio'".

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't think we need a TODO to add the right default. This is an async REST client so we're setting the right default value i.e. rest_asyncio.

What we may instead want to do is update the name / docstring of the class to make it clear that this is specific to async REST.

In future, we may want to spend some time to investigate if we can combine different transport and client class for gRPC and REST that we maintain.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So if this is an async rest client, why do we have a transport parameter at all? What do we do if the transport is not an async rest transport (ours or user-defined)? Might be worth filing a follow-up issue.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Users can still configure an instance of an async transport class but I agree that it is a bit weird. Can add a note about it in the issue that i file.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Filed: #726

try:
from google.auth.aio import credentials as ga_credentials # type: ignore
except ImportError as e: # pragma: NO COVER
raise ImportError(
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The name of the file and the name of the class do not make it clear that this is REST-specific. So async should work if at least one of { (REST dependencies), (gRPC dependencies) } are available. So it seems to me we should check for both sets of dependencies being absent before we error.

That, at least, when we get to the steady state of having everything implemented and cleaned up. If that's the reason we're not doing that now, let's add a TODO with a link to a tracking issue.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You're right. The name doesn't specify it but it is indeed REST specific. I agree that it's confusing. The name used here is consistent to what we have for the sync REST client / transport. We can have a discussion on what we want to name our async client / transport.

Updating the sync client name would be a breaking change.

page_token: Optional[str] = None,
timeout: Optional[float] = None,
metadata: Sequence[Tuple[str, str]] = (),
) -> pagers.ListOperationsAsyncPager:
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

AbstractOperationsClient has a compression parameter. Since This abstract class is, I assume, also applicable to gRPC, we should accept that parameter.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

AbstractOperationsClient shouldn't have a compression parameter since it's a REST only client. We have a separate client for gRPC.

@ohmayr ohmayr changed the base branch from async-rest-lro-support-in-core to main October 7, 2024 02:10
@ohmayr ohmayr changed the base branch from main to async-rest-lro-support-in-core October 7, 2024 05:10
@ohmayr
Copy link
Contributor

ohmayr commented Oct 7, 2024

@vchudnov-g I've addressed all of your comments.

The only thing that's left and is causing a bit of confusion is the name of client/transport class. Let's discuss this, address it accordingly, and get this over the line since I believe it's not a big change.

Copy link
Contributor

@vchudnov-g vchudnov-g left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, let's have a clearer naming convention for files and classes. We want to cover {sync, async} × {REST, gRPC}.

In the ideal world, we would have prefixes for each of those two dimensions:

Foo (eg Transport or Client)
SyncFooAsyncFoo
gRPCSyncFooRestSyncFoo gRPCAsyncFooRestAsyncFoo

Given historical naming and wanting to keep backward-compatibility, maybe we want something like this?:

BaseFoo (eg BaseTransport or BaseClient)
BaseSyncFooBaseAsyncFoo
FooRestSyncFoo AsyncFooRestAsyncFoo

Let's discuss the right names.....

Also, we should consider that we could have the "ideal" names in their own logical namespace, and assign the legacy names in the legacy namespace, e.e.:
transports.Foo = transports.ideal.gRPCSyncFoo
transports.AsyncFoo = transports.ideal.gRPCAsyncFoo
We'd have to figure out what name to use; it obviously wouldn't literally be "ideal".
Luckily, we can do this now or later, so we have options.

tests/unit/operations_v1/test_operations_rest_client.py Outdated Show resolved Hide resolved
Copy link
Contributor

@vchudnov-g vchudnov-g left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the discussion on the class hierarchy now and how we want it to evolve. Googlers, see b/372066378

super().__init__(
credentials=credentials, # type: ignore
# NOTE: If a transport is not provided, we force the client to use the async
# REST transport, as it should.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

So if this is an async rest client, why do we have a transport parameter at all? What do we do if the transport is not an async rest transport (ours or user-defined)? Might be worth filing a follow-up issue.

@ohmayr ohmayr merged commit 043785e into async-rest-lro-support-in-core Oct 7, 2024
5 checks passed
@ohmayr ohmayr deleted the operations-rest-async-client branch October 7, 2024 22:34
ohmayr added a commit that referenced this pull request Oct 7, 2024
* feat: implement `OperationsRestAsyncTransport` to support long running operations (#700)

* feat: Add OperationsRestAsyncTransport to support long running operations

* update TODO comment

* update TODO comment

* address feedback

* address feedback

* 🦉 Updates from OwlBot post-processor

See https://github.com/googleapis/repo-automation-bots/blob/main/packages/owl-bot/README.md

* fix mypy and lint issues

* minor fix

* add no cover

* fix no cover tag

* link coverage issue

* silence coverage issue

* fix statement name error

* address PR feedback

* address PR feedback

* address PR comments

---------

Co-authored-by: ohmayr <[email protected]>
Co-authored-by: Owl Bot <gcf-owl-bot[bot]@users.noreply.github.com>

* feat: implement async client for LROs (#707)

* feat: implement `AbstractOperationsAsyncClient` to support long running operations

* remove coverage guards

* address presubmit failures

* fix coverage for cancel operation

* tests cleanup

* fix incorrect tests

* file bugs

* add auth import

* address PR comments

* address PR comments

* fix unit tests and address more comments

* disable retry parameter

* add retry parameter

* address PR comments

---------

Co-authored-by: ohmayr <[email protected]>
Co-authored-by: ohmayr <[email protected]>

* 🦉 Updates from OwlBot post-processor

See https://github.com/googleapis/repo-automation-bots/blob/main/packages/owl-bot/README.md

---------

Co-authored-by: Anthonios Partheniou <[email protected]>
Co-authored-by: Owl Bot <gcf-owl-bot[bot]@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
size: xl Pull request size is extra large.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants