fix: handle null output_text content

[nightcityblade]

Fixes #3063

Summary

• allow ResponseOutputText.text to be null in typed models
• skip null text entries when aggregating response.output_text
• treat null text as unparsable in response parsing instead of raising

Testing

• python -m pytest -o addopts="" tests/lib/responses/test_responses.py -q

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3518e6a86d

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".

[chatgpt-codex-connector]

Guard streamed text concatenation after making text nullable

Changing ResponseOutputText.text to nullable allows response.content_part.added events to carry text: null, but ResponseStreamState.accumulate_event still does content.text += event.delta (src/openai/lib/streaming/responses/_responses.py:349-355). In streams where a content part starts with null and later receives response.output_text.delta, this will raise a TypeError and terminate the stream; initializing null text to "" (or branching before concatenation) avoids the crash.

Useful? React with 👍 / 👎.