Developer

Benchmarking

benchmarking.utils.compute_knn_graph_connectivities_and_distances(adata)

Compute approximate k-nearest-neighbors graph.

Data

data.initialize_dataloaders(node_masked_data)	Initialize edge-level and node-level training and validation dataloaders.
data.edge_level_split(data, edge_label_adj)	Split a PyG Data object into training, validation and test PyG Data objects using an edge-level split.
data.node_level_split_mask(data[, ...])	Split data on node-level into training, validation and test sets by adding node-level masks (train_mask, val_mask, test_mask) to the PyG Data object.
data.prepare_data(adata, ...[, adata_atac, ...])	Prepare data for model training including edge-level and node-level train, validation, and test splits.
data.SpatialAnnTorchDataset(adata, ...[, ...])	Spatially annotated torch dataset class to extract node features, node labels, adjacency matrix and edge indices in a standardized format from an AnnData object.

Models

models.utils.load_saved_files(dir_path, ...)	Helper to load saved model files.
models.utils.validate_var_names(adata, ...)	Helper to validate variable names.
models.utils.initialize_model(cls, adata, ...)	Helper to initialize a model.

Modules

modules.VGPGAE(n_input, n_fc_layers_encoder, ...)	Variational Gene Program Graph Autoencoder class.
modules.VGAEModuleMixin()	VGAE module mix in class containing universal VGAE module functionalities.
modules.BaseModuleMixin()	Base module mix in class containing universal module functionalities.
modules.compute_cat_covariates_contrastive_loss(...)	Compute categorical covariates contrastive weighted binary cross entropy loss with logits.
modules.compute_edge_recon_loss(...[, edge_incl])	Compute edge reconstruction weighted binary cross entropy loss with logits using ground truth edge labels and predicted edge logits.
modules.compute_gp_group_lasso_reg_loss(model)	Compute group lasso regularization loss for the masked decoder layer weights to enforce gene program sparsity (each gene program is a group; groups are normalized by the number of non-masked weights per group).
modules.compute_gp_l1_reg_loss(model, gp_type)	Compute L1 regularization loss for the rna decoder weights of gene programs of the type ´gp_type´ to encourage gene sparsity of those gene programs.
modules.compute_kl_reg_loss(mu, logstd)	Compute Kullback-Leibler divergence as per Kingma, D.
modules.compute_omics_recon_nb_loss(x, mu, theta)	Compute omics reconstruction loss according to a negative binomial model, which is often used to model omics count data such as scRNA-seq or scATAC-seq data.

NN

nn.OneHopAttentionNodeLabelAggregator(...[, ...])	One-hop Attention Node Label Aggregator class that uses a weighted sum of the omics features of a node's 1-hop neighbors to build an aggregated neighbor omics feature vector for a node.
nn.OneHopGCNNormNodeLabelAggregator(modality)	One-hop GCN Norm Node Label Aggregator class that uses a symmetrically normalized sum of the omics feature vector of a node's 1-hop neighbors to build an aggregated neighbor omics feature vector for a node.
nn.OneHopSumNodeLabelAggregator(modality)	One-hop Sum Node Label Aggregator class that sums up the omics features of a node's 1-hop neighbors to build an aggregated neighbor omics feature vector for a node.
nn.CosineSimGraphDecoder([dropout_rate])	Cosine similarity graph decoder class.
nn.FCOmicsFeatureDecoder(modality, entity, ...)	Fully connected omics feature decoder class.
nn.MaskedOmicsFeatureDecoder(modality, ...)	Masked omics feature decoder class.
nn.Encoder(n_input, ...[, n_addon_latent, ...])	Encoder class.
nn.MaskedLinear(n_input, n_output, mask[, bias])	Masked linear class.
nn.AddOnMaskedLayer(n_input, n_output, mask, ...)	Add-on masked layer class.

Train

train.Trainer(adata, model[, adata_atac, ...])	Trainer class.
train.eval_metrics(edge_recon_probs, edge_labels)	Get the evaluation metrics for a (balanced) sample of positive and negative edges and a sample of nodes.
train.plot_eval_metrics(eval_dict)	Plot evaluation metrics.