context_management=[{"type": "type"}] in tests, not 'compaction'

[qubez]

Confirm this is an issue with the Python library and not an underlying OpenAI API

• This is an issue with the Python library

Describe the bug

src/openai/types/responses/response_create_params.py gives us the supported string:

"""The context management entry type. Currently only 'compaction' is supported."""

It is sufficent to say that tests/api_resources/test_responses.py would fail without the right value for that key, or that the API actually able to run that test is not validating the one enum that would make something happen.

To Reproduce

1. look at the code
2. be astonished.

Code snippets

response = client.responses.create(
            background=True,
            context_management=[
                {
                    "type": "type",  # wot?
                    "compact_threshold": 1000,
                }
            ],

OS

AnyOS

Python version

Python 3.12

Library version

v2.20.0

[nomiveritas]

type": "compaction" the only supported enum
"compact_threshold" optional, but useful for token limit management
Works both with and without Pydantic validation
The API actually executes the context management logic when these parameters are used

"type": "compaction",
"compact_threshold": 1000
}

Best regards,
Nomiveritas

[nightcityblade]

I'd like to fix this. The test uses "type" as a placeholder value but it should be "compaction" per the API spec. Will submit a PR.

[nightcityblade]

After looking into this more, I see the test file is auto-generated by Stainless from the OpenAPI spec. The fix would need to be in the spec/generator rather than the test file directly. Unclaiming this.