Adds support for loading AWS bearer tokens via session constructor params

[jacoblee93]

Predicated on: boto/botocore#3671

Issue #, if available:

#4723

Description of changes:

Add support for aws_bearer_token as an explicit parameter on Session, Session.client(), and Session.resource(), enabling per-session and per-client bearer token authentication for services that support it (e.g. Bedrock).

Motivation

Currently, bearer tokens for services like Bedrock can only be configured via environment variables (e.g. AWS_BEARER_TOKEN_BEDROCK). This is a problem for multi-tenant systems that need different tokens for different clients within the same process — the environment variable is global and not thread-safe.

This PR adds an explicit aws_bearer_token parameter that follows the same pattern as aws_access_key_id and aws_secret_access_key: it can be set at the session level or overridden per-client.

Changes

• Added aws_bearer_token parameter to Session.__init__(), Session.client(), and Session.resource()
• Client-level token takes precedence over session-level token, which takes precedence over environment variables
• Uses is not None for precedence checks so that empty string is treated as an explicit value
• Includes a version guard that omits the parameter when None, for compatibility with older botocore versions that don't support it (same pattern as aws_account_id)

Depends on: corresponding botocore PR that adds aws_bearer_token to Session.create_client()

Usage

import boto3

# Session-level token
session = boto3.Session(aws_bearer_token='my-api-key')
client = session.client('bedrock-runtime', region_name='us-east-1')

# Per-client override
session = boto3.Session()
client = session.client('bedrock-runtime', aws_bearer_token='my-api-key')

# Multiple clients with different tokens
client_a = session.client('bedrock-runtime', aws_bearer_token='tenant-a-key')
client_b = session.client('bedrock-runtime', aws_bearer_token='tenant-b-key')

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.