[Copilot]

• Add Pipeline in /home/runner/work/PreToolkit/PreToolkit/src/preprocessing_toolkit/pipeline.py as a BasePreprocessor subclass with lazy validation in fit
• Override nested get_params / set_params behavior (steps, step replacement, <name>__<param>) with atomic error handling
• Implement chaining for fit, transform, and get_feature_names_out, including fit-failure unfitted behavior
• Export Pipeline from /home/runner/work/PreToolkit/PreToolkit/src/preprocessing_toolkit/__init__.py
• Add focused tests in /home/runner/work/PreToolkit/PreToolkit/tests/test_pipeline.py covering naming, validation, chaining, sklearn clone/grid-search/cross-val integration, and failure semantics
• Run targeted tests first, then ruff check, ruff format --check, and full pytest
• Run secret scanning on changed files and final parallel validation

• Fixes Add Pipeline with nested get_params / set_params #5