Add Ollama model name mapping for gpt-oss:20b

[kndtran]

Summary

• Add gpt-oss:20b -> gpt-oss_20b to BASE_MODEL_TO_CANONICAL_NAME in constants.py
• Ollama uses colons in model names (e.g. gpt-oss:20b), but HF repo paths use underscores (gpt-oss_20b)
• The mapping for granite4:micro -> granite4_micro already existed, but the equivalent for gpt-oss:20b was missing
• Without this fix, obtain_io_yaml fails with "not found" when using the Ollama model name

Test plan

• Verify obtain_io_yaml("answerability", "gpt-oss:20b", "ibm-granite/granite-lib-rag-gpt-oss-r1.0") resolves correctly

🤖 Generated with Claude Code

[kndtran]

Ok, turns out Ollama has a new engine that is required by but not yet supporting gpt-oss LoRAs. Claude found the issue in the Ollama code:

Ollama has two inference runners — the older llamarunner (C++ based, supports LoRA) and the newer ollamarunner (Go based, LoRA is a TODO). Models like gpt-oss, deepseek2, gemma3, qwen3, llama4, etc. are hardcoded to require the
ollamarunner via OllamaEngineRequired(). So LoRA adapters cannot be used with any of these models. The code has TODO(jessegross): LoRA loading but no issue or PR tracking it.