feat: add v2 -> v3 runtime migration warnings

[jam-jee]

Issue #, if available: N/A

Description of changes:

Adds visible, suppressible runtime warnings that guide SageMaker Python SDK v2 users toward v3, now that the v2 line is on the path of deprecation. This targets the master-v2 maintenance branch.

Two tiers, both built on a new warn_v2_deprecation() helper in sagemaker.deprecations:

• Tier 1 (package level): a single nudge emitted on import sagemaker, so every v2 user sees the migration notice regardless of which API they use.
  Sample code and output:

Code:  import sagemaker 
Output : SageMakerV2DeprecationWarning: You are using the SageMaker Python SDK v2, which is on path of deprecation. v3 is the actively developed major version.
See https://github.com/aws/sagemaker-python-sdk/blob/master/migration.md for the migration guide. Set SAGEMAKER_SUPPRESS_V2_WARNING=1 to silence this warning.

• Tier 2 (per interface): a targeted warning at the constructor of each major v2 interface that names the exact v3 replacement (plus a copy-pasteable import where the migration guide publishes one). Base-class call sites use type(self).__name__, so subclasses report their own name.

Interfaces covered by Tier 2: Estimator/framework estimators, Model, Predictor, Transformer, HyperparameterTuner, Processor, Pipeline, ModelMonitor, AutoML/AutoMLV2, Experiment, Run, ModelCard, AsyncPredictor, and the MLflow integration.

Sample code and output:

Code:  
from sagemaker.transformer import Transformer
Transformer(model_name="demo-model" , ***)
---
Output : /sagemaker/transformer.py:115: SageMakerV2DeprecationWarning: Transformer is part of the SageMaker Python SDK v2, which is on path of deprecation. In v3, use `sagemaker.core.shapes.TransformJob`.
See https://github.com/aws/sagemaker-python-sdk/blob/master/migration.md for the migration guide. Set SAGEMAKER_SUPPRESS_V2_WARNING=1 to silence this warning.

Design notes:

• SageMakerV2DeprecationWarning subclasses FutureWarning so it is shown by default (DeprecationWarning is silenced outside __main__).
• Fires at most once per feature per process; also emitted via logger.warning for log-captured environments (e.g. training jobs).
• Suppressible with SAGEMAKER_SUPPRESS_V2_WARNING=1 or standard Python warning filters.
• Workflow Step.__init__ is intentionally not instrumented because it breaks existing len(w) warning-count assertions; Pipeline covers that surface.

Testing done:

• Added unit tests in tests/unit/test_deprecations.py covering: default visibility (FutureWarning), once-per-process dedup, SAGEMAKER_SUPPRESS_V2_WARNING suppression, subclass naming, and exact-import rendering. 24/24 pass.
• Ran each touched module's unit suite in isolation (estimator, model, predictor, transformer, tuner, processing, pipeline, model_monitoring, automl, experiments, model_card, predictor_async, mlflow) — green except pre-existing, environment-related failures (missing optional mlflow/sklearn, offline botocore endpoints.json, a flaky timestamp-collision monitor test), each confirmed failing identically on the unmodified base via git stash.
• flake8 and black --line-length 100 clean on all changed files.

Merge Checklist

General

• I have read the CONTRIBUTING doc
• I certify that the changes I am introducing will be backward compatible, and I have discussed concerns about this, if any, with the Python SDK team
  • Note: changes are additive (new warnings only); no API signatures or behavior change. Targeted at the v2 maintenance line. Flagging for team confirmation that emitting a default-visible FutureWarning on import is acceptable messaging.
• I used the commit message format described in CONTRIBUTING
• I have passed the region in to all S3 and STS clients that I've initialized as part of this change. (N/A — no new clients initialized)
• I have updated any necessary documentation, including READMEs and API docs (if appropriate)

Tests

• I have added tests that prove my fix is effective or that my feature works (if appropriate)
• I have added unit and/or integration tests as appropriate to ensure backward compatibility of the changes
• I have checked that my tests are not configured for a specific region or account (if appropriate)
• I have used unique_name_from_base to create resource names in integ tests (if appropriate) (N/A — unit tests only)
• If adding any dependency in requirements.txt files, I have spell checked and ensured they exist in PyPi (N/A — no new dependencies)

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.