v0.2.0

• both synthetic and templated modes will no longer use Dirichlet distributions but rather Gaussian copulas, informed by Ledoit-Wolf-shrunk covariances and empirical marginals estimated from class-pooled data from glycowork (synthetic) or specific datasets (templated)
• missing data simulation has been switched from a power function to a logistic function
• batch effect simulation can now also simulate multiplicative batch effect (in addition to additive batch effects)
• simulate_paired workflow has been added to simulate same-sample multi-glycomics data with tunable coupling strength (e.g., N- and O-glycomics data of the same samples)