Large Language Models and Thematic Analysis: Human-AI Synergy in Researching Hate Speech on Social Media