API Reference

Note

API documentation is under construction. Current documentation is focused on core functionality.

Core Functionality

These modules provide functionality for utilizing SCimilarity embeddings for a variety of tasks, including cell type annotation, cell queries, and gene attribution scoring.

• scimilarity.cell_annotation

• scimilarity.cell_embedding

• scimilarity.cell_query

• scimilarity.cell_search_knn

• scimilarity.interpreter

Model Training

Training new SCimilarity models requires aggregated and curated training data. This relies on specialized data loaders that are optimized for random cell access across datasets, specialized variations of metric learning loss functions, and procedures for cell ontology aware triplet mining. The following modules include support for these training tasks.

• scimilarity.anndata_data_models

• scimilarity.nn_models

• scimilarity.training_models

• scimilarity.triplet_selector

• scimilarity.zarr_data_models

• scimilarity.zarr_dataset

Utilities

SCimilarity uses specific visualizations, ontology interfaces, and data preprocessing steps. These modules provide functionality useful for model training as well as a variety of SCimilarity analyses.

• scimilarity.ontologies

• scimilarity.utils

• scimilarity.visualizations