This is the first release of `modish`, a modality estimator. To estimate modalities, there are three steps:
1. Initialize a `ModalityEstimator`
2. Calculate log2 Bayes factors for each splicing event
3. Assign modalities based on the largest Bayes Factor score (and log2 bayes factor cutoff, default 3)
estimator = modish.ModalityEstimator()
log2_bayes_factors = estimator.fit_transform(psi_filtered)
modality_assignments = estimator.assign_modalities(log2_bayes_factors)
To see the parameterized family created from the estimator, do
import seaborn as sns
estimator = modish.ModalityEstimator()
`n` is the number of random variables to generate. Larger will result in smoother plots.
fig = estimator.violinplot(n=10000)
fig.tight_layout()
for ax in fig.axes:
ax.set(yticks=[0, 0.5, 1])
sns.despine()
