[DOCS][ML] Update documentation for default inference endpoints in semantic_text tutorials

[kosabogi]

Overview

With the introduction of default inference endpoints, manual creation of inference endpoints is no longer required for semantic search when using semantic_text.

The default (or preconfigured) endpoints start with a . character, we have 2 of them now:

• .elser-2-elasticsearch
• .multilingual-e5-small-elasticsearch

There is a new default value for the inference_id field (.elser-2-elasticsearch).

Update the following tutorials to align with the new functionality:

Required Updates:

• Tutorial: semantic search with semantic_text : Remove the section on manually creating inference endpoints, add information on default inference_id field.
• Tutorial: hybrid search with semantic_text: Remove the section on manually creating inference endpoints, add information on default inference_id field.
• Semantic text field type: Update with the new default value for the inference_id field.
• Get Inference API: Update code examples to include the new preconfigured endpoints.
• Ensure any references to configuring threads or vCPU counts are updated to reflect the integration of adaptive allocations.

[elasticsearchmachine]

Pinging @elastic/ml-core (Team:ML)