feat: add config validation step for model alias references

[lipikaramaswamy]

Priority Level

Medium

Task Summary

Add an Anonymizer.validate_config() method (similar to DD's DataDesigner.validate()) that checks model alias references, required fields, and config consistency before running the pipeline. Currently alias validation only happens at runtime inside load_workflow_config().

Technical Details & Implementation Plan

New public method on Anonymizer that validates:

• selected_models aliases exist in the model pool
• Required model aliases are present for the chosen replace strategy (e.g. LLMReplace needs a replacement_generator)
• Config fields are consistent

See TODO in model_loader.py:156.

Dependencies

No response