Text this: Creating a Novel Attention-Enhanced Framework for Video-Based Action Quality Assessment