fix(ai): response stream chunking de-buffering

[oowl]

Summary

Fixes an issue where "AI streaming responses" were returning all inside a single chunk, instead of being returned chunk-by-chunk to the client.

Also fixes other parsing issues with Bedrock, where the wrong response content-type was used.

Checklist

• The Pull Request has tests
• A changelog file has been created under changelog/unreleased/kong or skip-changelog label added on PR if changelog is unnecessary. README.md
• There is a user-facing docs PR against https://github.com/Kong/docs.konghq.com - bugfix, no new docs

Issue reference

FTI-6419

[tysoekong]

Yep approved as discussed, I've tested all dependent AI services manually and the automation will handle the edge cases.