Add OCI Generative AI provider support

Enhancement summary

Add support for Oracle Cloud Infrastructure (OCI) Generative AI to aisuite through OCI's OpenAI-compatible Chat Completions endpoint.

Motivation

OCI users should be able to access OCI-hosted models through aisuite’s provider-agnostic, OpenAI-shaped API, without introducing provider-specific application code.

Proposed behavior

Introduce an oci provider supporting:

• API-key authentication for local development and testing.
• Resource Principal authentication for OCI-managed production workloads.
• Synchronous and asynchronous chat completions.
• Streaming chat completions.
• Model identifiers in the form oci:<model>, for example:

model="oci:openai.gpt-5.5"

Configuration should be supplied through provider settings or the following environment variables: OCI_GENAI_REGION OCI_GENAI_API_KEY OCI_GENAI_PROJECT (optional)

Documentation and testing

The change includes: OCI usage and authentication documentation. Mocked unit tests for provider configuration and request behavior. Opt-in live integration tests for sync, async, and streaming calls.

Pull request

PR #397