generate_sim_data: Generate data for running simulations with mmBPFA
In EandrewJones/mmBPFA: Multi-Modal Beta-Process Factor Analysis

generate_sim_data produces simulated data that can be used for testing the mmBPFA sampler prior to production runs to ensure it will behave as expected.

generate_sim_data(
  mode = "fixed",
  n_levels = 2,
  n = 1000,
  p = 500,
  K_true = 10,
  sparsity = NULL,
  a = 10,
  b = 1,
  d = 2,
  e = 1,
  prop_missing = 0,
  noise = 0.05,
  seed = 123
)

`mode`	Sets the type of margins for the simulated data. Must be one of "fixed", "multi" or "mixed." Default is "fixed."
`n_levels`	Max number of unique values per margin. When mode is "fixed" n_levels is constant across margins. When mode is "multi" a margin may take on unique values up to n_levels. When mode is "mixed", n_levels determines max number of arbitrary distributions data may be drawn from. Default is 2.
`n`	Number of observations. Default is 1000.
`p`	Number of features. Default is 500.
`K_true`	True number of latent dimensions. Default is 10.
`sparsity`	Numeric value between 0 and 1 that determines amount of sparsity in the factor loadings. If sparsity argument is set, zeros are uniformly distributed across all dimensions. Default is NULL.
`a`	A hyperparameter determining the distribution of sparsity across the factor loadings. Default is 10
`b`	A hyperparameter determining the distribution of sparsity across the factor loadings. Default is 1.
`d`	A hyperparameter determining the shape of the factor precisions (gamma_k). Default is 2.
`e`	A hyperparameter determining the scale of the factor precisions (gamma_k). Default is 1.
`prop_missing`	Numeric value between 0 and 1 determining the proportion of missingness in the data. Default is 0.
`noise`	Numeric value between 0 and 1 determining how much noise should be induced into the data. Default is 0.05.
`seed`	Random seed. Default is 123.