fix(ChatGptController): Chinese encoding issue

[Trioc]

Summary

This pull request fixes an encoding issue in the AI component where Chinese responses
returned from the OpenAI API were displayed as "??????????" in the WISE dialog.

Root Cause

The OpenAI HTTP response was being read using the default character set, which resulted
in the response being interpreted as ISO-8859-1. Since Chinese characters cannot be
represented in ISO-8859-1, the UI displayed garbled text ("???????").

Fix Implemented

• Updated ChatGptController.java to explicitly read the HTTP response stream using
  ISO-8859-1 when parsing the OpenAI response.
• This ensures that the returned JSON string remains intact, and the Chinese content is
  correctly handled by the downstream components in WISE.

Why ISO-8859-1?

• OpenAI’s API sometimes returns response headers indicating ISO-8859-1.
• Using ISO-8859-1 prevents character loss during the initial read operation because
  it maps bytes directly (1:1), allowing the actual UTF-8 content inside the JSON to be
  preserved for correct later parsing.

Impact

• Chinese character responses are now displayed correctly in the WISE AI dialog.
• No behavior changes for English or other ASCII-only responses.
• The fix is isolated to the AI component and does not affect other subsystems.

Testing

• Verified locally in our WISE deployment that Chinese responses from OpenAI are now
  displayed correctly in the dialog.
• Confirmed no regression for English responses.

[hirokiterashima]

Thanks for submitting this PR. Fix works as described. Can you clean up ChatGptcontroller.java (I added suggestions inline), and remove wise_api_local_changes.patch file?

[Trioc]

Thanks for the review! I’ve removed the debug print, cleaned up the unused imports, and deleted the local patch file as suggested. The PR has been updated.

[hirokiterashima]

🎉 This PR is included in version 1.13.1 🎉

The release is available on GitHub release

Your semantic-release bot 📦🚀