feat: use llama.cpp server

[vansangpfiev]

Describe Your Changes

• Ready to review, e2e tests work on VMs, still checking why they do not work on CIs

This pull request includes various updates and improvements across multiple files, focusing on version updates, code refactoring, and adding new functionality. The most important changes include updating version strings, refactoring code to use constants, and adding new engine support.

Version Updates:

• Updated version strings in docs/static/openapi/cortex.json to reflect new versions and naming conventions. [1] [2] [3] [4] [5]

Code Refactoring:

• Removed the IsSupported method from EngineI class in docs/docs/engines/engine-extension.mdx and engine/cortex-common/EngineI.h for cleaner interface. [1] [2]
• Replaced hardcoded organization and repository names with constants in various files for maintainability. [1] [2] [3] [4]

New Functionality:

• Added support for local-engine in CMakeLists.txt files to include new engine extension. [1] [2]
• Added a custom command for MSVC builds in engine/CMakeLists.txt to copy necessary files post-build.

Logging and Debugging:

• Added logging statements in engine/controllers/engines.cc for better traceability of engine versions and variants during installation. [1] [2]

Test Updates:

• Updated test cases in engine/e2e-test/api/engines to use new version strings and variant names. [1] [2] [3] [4] [5] [6] [7] [8]

Fixes Issues

• Replace cortex.llamacpp with minimalist fork of llama.cpp #1728

Self Checklist

• Added relevant comments, esp in complex areas
• Updated docs (for bug fixes / features)
• Created issues for follow-up changes or refactoring needed

[github-actions]

Preview URL: https://be4cfec3.cortex-docs.pages.dev