Integrating Large Language Models and Knowledge Graphs for Extraction and Validation of Textual Test Data