Add support for DEFAULT_LLM_MODEL environment variable

[Abbeyme1]

Summary

This PR introduces support for a new environment variable DEFAULT_LLM_MODEL to configure the default LLM model used by the web-ui. This eliminates the need to repeatedly set the model manually.

Changes Made

• Added support for DEFAULT_LLM_MODEL in the executable script.

• Updated code to fetch the model using:

  os.getenv("DEFAULT_LLM_MODEL", config.model_names[os.getenv("DEFAULT_LLM", "openai")][0])

• Ensures a consistent and convenient way to define the default model through environment variables.

---

How to Test

1. Set the environment variable before running:

   export DEFAULT_LLM_MODEL=gpt-4

2. Start the web-ui

3. The selected model should default to the value set in DEFAULT_LLM_MODEL without requiring manual selection.

---

Notes

• If DEFAULT_LLM_MODEL is not set, fallback logic still uses the default model from config.model_names based on the DEFAULT_LLM provider or defaults to "openai".

---

Summary by cubic

Added support for the DEFAULT_LLM_MODEL environment variable to set the default LLM model in the web UI.

• New Features
  • The web UI now uses DEFAULT_LLM_MODEL to pick the default model, so users no longer need to select it manually.
  • If DEFAULT_LLM_MODEL is not set, the app falls back to the existing default logic.

[CLAassistant]

All committers have signed the CLA.

[cubic-dev-ai]

cubic reviewed 2 files and found no issues. Review PR in cubic.dev.