Parsimonious Contaminated Shifted Asymmetric Laplace Mixtures: Unsupervised Learning with Outlier Identification for Asymmetric Clusters in High Dimensions