refactor llm info

[pelikhan]

Move LLMs inormation to own file

---

Pull Request Description

Key Changes

• 🔧 Dependency Updates:

  • Updated several dependencies across package.json files, including @astrojs/starlight, astro, @inquirer/prompts, @huggingface/transformers, and esbuild, among others, to newer versions for improved stability and features.

• 🆕 Model Aliases Enhancements:

  • Added support for defining and managing model aliases in llms.json and across CLI commands, enabling flexible alias-based model selection for large, small, vision, and reasoning models.
  • Introduced a new info models alias CLI command to display model alias mappings.

• 🛠️ CLI Improvements:

  • Enhanced CLI commands (run, test, etc.) with additional model provider options and streamlined alias handling.
  • Refactored and modularized code for better maintainability (e.g., replaced modealias.ts with modelalias.ts).

• 📚 Documentation Updates:

  • Fixed a typo in provider usage (mistal → mistral).
  • Expanded schema definitions in llms.json to include aliases and updated documentation to reflect new alias capabilities.

• 🧩 Feature Additions:

  • Added a new judgeClassify utility in the sample project for task classification based on predefined categories, showcasing advanced prompt usage.

• ⚙️ Code Refactoring:

  • Consolidated default model configurations into a reusable function (defaultModelConfigurations) for consistency across runtime environments.
  • Improved handling of empty values and undefined properties in utility functions.

• 🧪 Testing Enhancements:

  • Updated test cases to validate new CLI commands and alias functionality.

• 🚀 Performance Improvements:

  • Optimized logit bias computation by reducing unnecessary warnings and improving verbosity control.

Summary

This PR introduces significant improvements to dependency management, CLI functionality, and model alias support, while enhancing documentation and testing. These changes streamline workflows, improve maintainability, and provide a more robust developer experience. 🚀

AI-generated content by prd may be incorrect

[github-actions]

LGTM 🚀

Changes look good! The change from logWarn to logVerbose is appropriate if the verbose logging level is suitable for this scenario. There are no functional issues or concerns noted in the provided diff. The change improves the log level consistency, which can be beneficial for better control and analysis of logs.

AI-generated content by pr-review may be incorrect

[github-actions]

Investigator report

• run failure
  , run last success
  , 0d394b9
  , diff f16d915...0d394b9

AI-generated content by gai may be incorrect