tangram.mapping_optimizer

Description

Library for instantiating and running the optimizer for Tangram. The optimizer comes in two flavors, which correspond to two different classes: - Mapper: optimizer without filtering (i.e., all single cells are mapped onto space). At the end, the learned mapping matrix M is returned. - MapperConstrained: optimizer with filtering (i.e., only a subset of single cells are mapped onto space). At the end, the learned mapping matrix M and the learned filter F are returned.

Classes

Mapper(S, G[, d, d_source, lambda_g1, …])

Allows instantiating and running the optimizer for Tangram, without filtering.

MapperConstrained(S, G, d[, lambda_d, …])

Allows instantiating and running the optimizer for Tangram, with filtering.