fix: ModelTrainer and HyperparameterTuner missing environment variables (5613)

[aviruthen]

Description

The HyperparameterTuner in V3 SDK does not propagate the ModelTrainer's environment attribute to the training job definition. The _build_training_job_definition method in tuner.py constructs the TrainingJobDefinition for the HPT API but completely omits the environment field from the model_trainer. The word 'environment' never appears in tuner.py. The fix requires adding environment to the training job definition built by the tuner, reading it from model_trainer.environment. This affects both the single-trainer path (_build_training_job_definition) and the multi-trainer path (_build_training_job_definitions). The environment should be included as a top-level field in the HyperParameterTrainingJobDefinition, which the SageMaker HPT API supports via the Environment key.

Related Issue

Related issue: 5613

Changes Made

• sagemaker-train/src/sagemaker/train/tuner.py
• sagemaker-train/tests/unit/train/test_tuner.py

AI-Generated PR

This PR was automatically generated by the PySDK Issue Agent.

• Confidence score: 85%
• Classification: type: bug
• SDK version target: V3

Merge Checklist

• Changes are backward compatible
• Commit message follows prefix: description format
• Unit tests added/updated
• Integration tests added (if applicable)
• Documentation updated (if applicable)