henribru / google-api-python-client-stubs Goto Github PK

Hi,
First, thanks for maintaining this library.
I'm using google-api-python-client v1.12.11 in tandem with Airflow 2.6.1's constraints file.

I'm trying to install google-api-python-client-stubs to match, but not finding any version that pip is happy with.

If I try google-api-python-client-stubs==1.12.0 I get,

google-api-python-client-stubs 1.12.0 depends on google-api-python-client<3.0.0 and >=2.64.0
The user requested (constraint) google-api-python-client==1.12.11

and if I just specify ``google-api-python-client-stubsinrequirements.txt` and let pip try to figure it out, it tries all the available releases and ends with,

ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

Please advise.
Thank you!

My understanding is that these types are auto-generator, can you perform an update and/or publish the generator so this can be done automatically?

E.g. list_next from the Content API

Could this library be extended to provide stubbing for tests?
I have recently started developing against GCP in python and I really miss the AWS SDK test stubs for Ruby.
If this feels like a good fit, I would be interested in collaborating.

Hi, thanks for the work on this library!

The google-api-python-client library makes available the HttpMock class for testing purposes (see https://github.com/googleapis/google-api-python-client/blob/master/docs/mocks.md#example)

I don't believe this is compatible with the signature defined at https://github.com/henribru/google-api-python-client-stubs/blob/master/googleapiclient-stubs/discovery.pyi#L346.

Would an | operator on the types be the change we want here or would you recommend doing something like:

http = Http(HttpMock("test_response.json", {"status": "200"}))

Sorry if I ask, and if I am is because I value this package so good job on that.

But really, what should I do in order to make this work? The documentation doesn't seem to explain much besides installing the package (which I did obviously).

I am using poetry and installed both googleapiclient and this package using poetry add. At first I tried installing the stubs in the DEV group, then I also tried in the main group while trying many thing to try and get this working.

So, I'm trying to use Google Calendar API, right?
Basically, if I don't import from googleapiclient._apis.calendar.v3.resources, I cannot use CalendarResource because it's undefined, but that import doesn't work! I feel like I'm missing something really simple here, please help. Thank you very much

https://cloud.google.com/api-keys/docs/reference/rest

Similar to #12

There are methods like get_media and watch_media that are missing from the stubs

https://github.com/googleapis/google-api-python-client/blob/d0f8a05b5c236f99d88fc6e0ef671a0a3eb58a4b/googleapiclient/discovery.py#L1471-L1479

See also:

• https://developers.google.com/resources/api-libraries/documentation/drive/v3/python/latest/drive_v3.files.html
• https://googleapis.github.io/google-api-python-client/docs/dyn/drive_v3.files.html
• https://github.com/googleapis/google-api-python-client/blob/e4faf093d059688eb1a57689c2ac79563b61eaf5/docs/dyn/drive_v3.files.html

The methods that google-api-python-client generates at runtime actually has docstrings. It would be useful to add these in the stubs so that they're shown in your IDE

Currently nested resources are defined as nested classes. This is kinda convenient because it means you only need to import the base resource, e.g. you can do from googleapiclient._apis.analytics.v3.resources import AnalyticsResource and then you can get the nested management resource as AnalyticsResource.ManagementResource. It also mirrors how you get that nested resource in your code, i.e. management_resource = analytics_resource.management(). The main issue with this is that it ends up polluting the autocomplete on the resources. For an example, notice how both management and ManagementResource show up here:

The question is if it could it be worth restructuring this in some way to avoid this issue