As a developer, I want to have a sliced OpenAPI document with only the endpoints and components used by my solution when generating a client

### Is your feature request related to a problem? Please describe the problem.

Kiota produces large OpenAPI documents even when only one endpoint is selected because it doesn’t remove unused components.

### Client library/SDK language

None

### Describe the solution you'd like

When generating clients, Kiota should be able to trim unused components from the original OpenAPI document to generate a clean sliced OpenAPI document with only the information that is important for my solution.

For clients:

Remove all unused components and keep all used components including schemas, parameters, responses.
Remove all extensions (starts with x-) that don't match with our internal documentation of supported extensions in Kiota.

### Additional context

_No response_

No file or test is named. Start by tracing the client-generation entry point from an OpenAPI input and identify where referenced components and extensions are collected. Done means the generated document retains only used schemas, parameters, and responses, removes unsupported x- extensions, and has coverage for the slicing behavior.

Written by the indexing model from the issue text.