fix: correct prompt_cache_retention literal from "in-memory" to "in_memory"#2923
fix: correct prompt_cache_retention literal from "in-memory" to "in_memory"#2923carlosinfantes wants to merge 2 commits intoopenai:mainfrom
Conversation
…emory" The SDK declares prompt_cache_retention with Literal["in-memory", "24h"] (hyphen), but the API rejects "in-memory" with a 400 and only accepts "in_memory" (underscore). Fixes openai#2883
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 16b7359d5b
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
| prompt: Optional[ResponsePromptParam] | Omit = omit, | ||
| prompt_cache_key: str | Omit = omit, | ||
| prompt_cache_retention: Optional[Literal["in-memory", "24h"]] | Omit = omit, | ||
| prompt_cache_retention: Optional[Literal["in_memory", "24h"]] | Omit = omit, |
There was a problem hiding this comment.
After changing the method/type signatures to Literal["in_memory", "24h"], the generated tests still pass "in-memory", which now fails static typing (pyright -p pyproject.toml reports argument-type errors at tests/api_resources/chat/test_completions.py:76,210,519,653 and tests/api_resources/test_responses.py:55,142,464,551). This introduces a typecheck regression in the repo state unless the test call sites are updated alongside the API surface change.
Useful? React with 👍 / 👎.
Update test call sites to match the corrected Literal type. Fixes openai#2883
1 participant
Summary
The SDK declares
prompt_cache_retentionwithLiteral["in-memory", "24h"](hyphen), but the API rejects"in-memory"with a 400 and only accepts"in_memory"(underscore). Using the SDK-typed value as-is always fails.Changes
Updated all occurrences of
Literal["in-memory", "24h"]toLiteral["in_memory", "24h"]across 7 files:src/openai/types/responses/response_create_params.pysrc/openai/types/responses/response.pysrc/openai/types/responses/responses_client_event.pysrc/openai/types/responses/responses_client_event_param.pysrc/openai/types/chat/completion_create_params.pysrc/openai/resources/responses/responses.pysrc/openai/resources/chat/completions/completions.pyFixes #2883