Fix mteb_to_codeweaver.py: undefined names, broken init block, wrong imports and sys.path

[Copilot]

scripts/model_data/mteb_to_codeweaver.py was partially implemented with several ruff F821 (undefined name) errors preventing it from running.

Fixes

• KNOWN_ALIASES undefined — RootJson.aliases computed property referenced KNOWN_ALIASES but it was never defined. Added definition built from OLLAMA_ALIASES:

  KNOWN_ALIASES: dict[str, dict[ModelName, ModelName]] = {
      "ollama": OLLAMA_ALIASES,
  }

• Init block accidentally stringified — The if/else block that defines _ROOT, DATA, MODEL_MAP_DATA, ALIAS_MAP_DATA, and FLATTENED_ALIASES was wrapped in triple quotes, making it a no-op string literal instead of executable code. Restored as proper if JSON_CACHE.exists() / else block. Also dropped stale SPARSE_MODELS reference — that field no longer exists on RootJson.

• Broken codeweaver imports — Both imports pointed to non-existent re-export locations:

  • from codeweaver.core import Provider → from codeweaver.providers.provider import Provider
  • from codeweaver.providers import PartialCapabilities → from codeweaver.providers.embedding.capabilities.types import PartialCapabilities

• Wrong sys.path.insert depth — Was inserting scripts/ (one level too shallow for the src/ layout), fixed to Path(__file__).parent.parent.parent / "src".

Supporting files

Created the scripts/model_data/ directory with the required data files the script depends on: secondary_providers.json, a minimal hf-models.json, and their REUSE license files.

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[Copilot]

Copilot reviewed 5 out of 9 changed files in this pull request and generated no comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.