Allow customization of baseURL for Gemini integration

[vehkiya]

Describe the feature

Add an optional config flow step that allows for customization of base URL for Gemini integration. Default remains unchanged and users only override it on their own risk if they know what they are doing.

Use cases

Allows to use Gemini in regions where Gemini and Vertex AI is available, but AI Studio is not.
In theory, would allow the usage with a self-hosted Gemma instance

Anything else?

This practice is becoming somewhat standard in the industry. The expected API and endpoints are unchanged. Platforms like n8n already implement it this way.

[vehkiya]

Vertex AI Studio (https://console.cloud.google.com/vertex-ai/studio) implements the same endpoints as the other AI Studio as described in https://ai.google.dev/api
Notably, it exposes the same generateContent and streamGenerateContent as well.