Debuging specific test case for JSON returned from AI

[mercury64]

I have an issue, and I've narrowed it down to the JSONParser, Pydantic, executor and CommandResult model.
JSON is being returned correclty, but the CommandResult model does not populate.

Here is my debug log:

2025-01-13 15:55:25,582 DEBUG [core.llm.ollama_client] *** OLLAMA RESPONSE ***
2025-01-13 15:55:25,582 DEBUG [core.llm.ollama_client] {
     "CommandResult": {
         "analysis": "The command was executed successfully without any errors.",
         "success": true
     }
}

2025-01-13 15:55:25,582 DEBUG [core.llm.parser] {'CommandResult': {'analysis': 'The command was executed successfully without any errors.', 'success': True}}
2025-01-13 15:55:25,582 DEBUG [core.llm.base] Error parsing LLM response: Invalid JSON format:
- `analysis`: missing (Field required)
- `success`: missing (Field required), asking LLM to retry
Traceback (most recent call last):
  File "/Users/vi1154/Programming/gpt-pilot/core/llm/parser.py", line 139, in __call__

How can I run a specific test case for this?

ollama_client is basically a copy of openai_client.

[mercury64]

This seems to be the most related bug #738