fix parse exception when sse stream returns incomplete data

[wliu6v]

I've noticed that the first response from SSE is sometimes incomplete, where it might split a single JSON object across two responses.

The first:
{"id":"chatcmpl-8NY6x7nqghZB9FgFciPtk5

The second:
azY6e6e","object":"chat.completion.chunk","created":1700623359,"model":"gpt-4-0613","choices":[{"index":0,"delta":{"role":"assistant","content":""},"finish_reason":null}]}

By catching a jsonDecode exception, we can identify this situation and attempt to concatenate the incomplete response with the next one to parse them together, which should solve the issue.

[codecov-commenter]

Codecov Report

Attention: 13 lines in your changes are missing coverage. Please review.

Comparison is base (a17b4fd) 83.87% compared to head (be02275) 83.87%.
Report is 1 commits behind head on main.

Files	Patch %	Lines
lib/src/client/openai_client.dart	13.33%	13 Missing ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #80   +/-   ##
=======================================
  Coverage   83.87%   83.87%           
=======================================
  Files          71       71           
  Lines        1048     1048           
=======================================
  Hits          879      879           
  Misses        169      169           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.