MedFLIP: Medical Vision-and-Language Self-supervised Fast Pre-Training with Masked Autoencoder