Add Ollama model name mappings for Granite 4.1 to intrinsics adapter resolution

[kndtran]

The intrinsics adapter resolver (download_intrinsic_adapter in mellea/formatters/granite/intrinsics/util.py) uses BASE_MODEL_TO_CANONICAL_NAME to normalize the base model name before looking up LoRA/aLoRA adapter paths in the Granite Intrinsics Library.

When serving intrinsics adapters via Ollama, the base model name is the Ollama identifier (e.g. granite4.1:3b), which needs to be mapped to the GGUF adapter directory name on HF (e.g. granite4.1_3b). The mapping for Granite 4.0 Micro already exists ("granite4:micro" → "granite4_micro"), but the Granite 4.1 equivalents are missing.

This causes download_intrinsic_adapter to fail with:

ValueError: Intrinsic 'answerability' as LoRA adapter on base model 'granite4.1:3b' not found...
Searched for path answerability/granite4.1:3b/lora/io.yaml

The resolver looks for granite4.1:3b literally in the path instead of normalizing it to granite4.1_3b.

Proposed fix

Add the following entries to BASE_MODEL_TO_CANONICAL_NAME in mellea/formatters/granite/intrinsics/constants.py:

    "granite4.1:3b": "granite4.1_3b",
    "granite4.1:8b": "granite4.1_8b",
    "granite4.1:30b": "granite4.1_30b",