UPDATES REGARDING CLI CONFIG

Google Summer of Code
1

BeagleMind CLI — Recent Update Summary

Hello mentors and BeagleBoard community,

I’m excited to share a progress update on the BeagleMind CLI, especially since the latest improvements have now been merged into the main repository.

Over the past few days, I focused on reorganizing and improving the internal structure of the CLI to make it more maintainable, scalable, and user-friendly. Below is a summary of what has been completed and merged.

1. Configuration System Refactor (Merged)

A major improvement was redesigning how the CLI handles configuration:

  • All backend and model lists (groq, openai, ollama) were moved from the CLI into the ConfigManager, where they logically belong.
  • Defaults such as backend, model, temperature, and collection name are now sourced from a single config file instead of being scattered across files.
  • The ConfigManager now provides cleaner helper methods like
    get_backends(), get_models(backend), and get_default_model().

This creates a more organized, future-proof configuration structure.

2. Removal of Hardcoded Arrays

Static arrays such as GROQ_MODELS, OPENAI_MODELS, and OLLAMA_MODELS have been fully removed from the CLI.

The CLI now retrieves all backend and model information dynamically via:

self.config_manager.get_models("<backend>")

This centralizes data management and avoids future duplication or drift.

3. Unified Validation & Error Handling

Backend and model validation now depends entirely on ConfigManager, eliminating inconsistencies caused by earlier hardcoded lists.
The CLI now provides clearer, more accurate error messages referencing dynamic config values.

4. Improved Model Display Table

The show-models command has been updated to:

  • Pull backend/model options from the config file
  • Correctly mark Ollama as Local
  • Display all backend categories in a clean, consistent table

This results in better clarity for users exploring available model options.

5. General Clean-Up and Quality Fixes

The PR also included:

  • Removal of redundant code and commented-out blocks
  • Fixes to get/set inconsistencies
  • Ensuring all fallback defaults come from ConfigManager
  • More consistent CLI formatting and structure

All mentor feedback has been addressed, and the final changes are now part of the main branch.

A Heartfelt Thank You

I want to express my sincere gratitude to @fayezzouari for the continuous guidance, patient reviews, and encouraging feedback throughout this process.
Your mentorship has been incredibly motivating and has helped me improve both my code quality and my understanding of good project structure.
Thank you for taking the time to review each change carefully and helping me grow as a contributor .I will be gratefull to work more with all of you.
THANK YOU