[codex] fix Codex image generation transport

[cyeinfpro]

What changed

This PR routes OpenAI OAuth image generation/edit requests through ChatGPT's dedicated Codex images endpoints instead of the Responses API image tool path.

• Use /backend-api/codex/images/generations and /backend-api/codex/images/edits for OAuth image accounts.
• Preserve native image request fields such as size, quality, background, output_format, output_compression, and n.
• Normalize Codex image responses back into the existing OpenAI Images API shape.
• Convert URL-only upstream image outputs to b64_json when requested by downstream clients.
• Preserve synthetic SSE compatibility for downstream stream: true callers.

Why

The previous OAuth image path wrapped image requests inside the Responses API image-generation tool. That path did not reliably preserve native gpt-image-2 behavior, especially large output sizes such as 3840x2160, and it made multi-image n handling depend on the Responses tool wrapper.

The dedicated Codex images endpoint accepts the native image payload directly, so gpt-image-2 2K/4K requests and n are passed to the upstream service without being translated into tool arguments.

Validation

• go test -tags unit ./internal/service -run 'TestOpenAIGatewayServiceForwardImages|TestOpenAIGatewayServiceParseOpenAIImagesRequest|TestBuildOpenAIImagesURL'
• Live deployment smoke tested separately with gpt-image-2 4K generation returning 3840x2160.

[github-actions]

Thank you for your contribution! Before we can merge this PR, we need you to sign our Contributor License Agreement (CLA).

To sign, please reply with the following comment:

I have read the CLA Document and I hereby sign the CLA

You only need to sign once — it will be valid for all your future contributions to this project.

---

I have read the CLA Document and I hereby sign the CLA

---

_{You can retrigger this bot by commenting recheck in this Pull Request. Posted by the CLA Assistant Lite bot.}