Fix model name extraction from Ollama API response

[ramasubbareddy1224]

Summary

This PR fixes the model name extraction logic in ensure_model_pulled() to correctly parse the Ollama API response structure.

Changes

• Line 30: Added extraction logic to parse model names from the ollama.list() response
  • The API returns: {'models': [{'name': 'llama3.2:1b', 'model': 'llama3.2:1b', ...}]}
  • Extracts the model field from each entry in the models array using: {m["model"] for m in available_model_list.get("models", [])}
  • Ensures the function can properly identify which models are already available locally

Testing

• Verified that the extraction correctly identifies available models from the Ollama API response
• Confirmed the function can now properly check if a model exists before attempting to pull it