Add LoRA Load and Unload APIs

[danehans]

Currently, the project loads all LoRA's specified by the config. To more closely align with LoRA support for vLLM, add /v1/load_lora_adapter and /v1/unload_lora_adapter API endpoints (xref).

[shmuelk]

These endpoints are already suported by the llm-d-inference-sim. See pkg/llm-d-inference-sim/simulator.go lines 154 and 155

We will add documentation indicating which other vLLM compatible HTTP endpoints are suported.

[shotarok]

@danehans I was wrong in kubernetes-sigs/gateway-api-inference-extension#980. The endpoints are implemented in 0549810, so I can work on an e2e test with the endpoints!