feat: support OpenAI model subpaths

[spduan78]

Summary

Add OpenAI model resource subpath support for /v1/models/{model} in the AI proxy.

Higress already recognizes the model list endpoint /v1/models, but the OpenAI Models API also includes retrieving a model and deleting a fine-tuned model through the model resource path. Without a dedicated capability, these resource requests cannot be classified or remapped by capability path.

Changes

• Add ApiNameRetrieveModel and PathOpenAIRetrieveModel.
• Add OpenAI default capability mapping for the model resource path.
• Add regex-based route detection for /v1/models/<model_id>.
• Add {model_id} placeholder substitution in MapRequestPathByCapability.
• Allow the new capability in provider JSON configuration.
• Add regression tests for API detection, path remapping, default capabilities, and config parsing.

The internal placeholder is named {model_id} to avoid colliding with Azure deployment paths that already use {model} as a deployment placeholder.

Reference: OpenAI Models API reference documents list models, retrieve model, and delete fine-tuned model endpoints.

Tests

GOCACHE=/private/tmp/higress-go-cache go test ./... -run 'Test_getApiName|TestProviderConfig_FromJsonModelCapabilities|TestOpenAIProviderInitializer_DefaultCapabilitiesModels|TestUtil' -count=1
GOCACHE=/private/tmp/higress-go-cache go test ./provider -run 'TestAzureDefaultOpenAIV1Capabilities|TestOpenAIProviderInitializer_DefaultCapabilitiesModels|TestProviderConfig_FromJsonModelCapabilities' -count=1
GOCACHE=/private/tmp/higress-go-cache go test . -run 'Test_getApiName|TestUtil' -count=1
GOCACHE=/private/tmp/higress-go-cache go test ./...

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}