report

besca.tl.report(adata_pred, celltype, method, analysis_name, train_datasets=[], test_dataset='', merge='', name_prediction='auto_annot', name_report='auto_annot', use_raw=False, genes_to_use='', remove_nonshared=False, clustering='leiden', asymmetric_matrix=True, results_folder='./', delimiter=',', verbose=False)

reports basic metrics, produces confusion matrices and plots umap of prediction

Writes out a csv file containing all accuracy and f1 scores. Writes normalized and absolute confusion matrices, as well as umap prediction comparisons to ./figures.

Parameters:

• adata_pred (AnnData) – original adata object with name_prediction column

• celltype (str) – celltype column on which the prediction was performed

• method (str) – method that was used for prediction.

• analysis_name (str) – name of the analyis, used for writing files

• train_datasets (list) – list of used training datasets

• test_dataset (str) – name of test dataset

• merge (str) – what merging was performed

• name_prediction ("auto_annot"| default = "auto_annot") – observation name containing the prediction to compare with.

• name_report ("auto_annot"| default = "auto_annot") – prefix of the report

• use_raw (bool | default = False) – if anndata.raw was used

• genes_to_use (list or string | default = ‘all’) – what geneset wsa used

• remove_nonshared (`bool`|default = False) –

• clustering (str | default = leiden) – clustering that was used in original analysis of testing set, needed for umap plotting

• asymmetric_matrix (bool | default = True) – if False returns square confusion matrix, if True it only shows possible combinations

• results_folder (str | default = ‘./’) – output directory. A figures folder will be generated within it.

• delimiter (str | default = ‘,’) – separator between fields in the csv/txt report file

• verbose (bool | default = False) – print verbose messages to standard out

Returns:

A matplotlib figure element containing the riveplot generated for interactive display.

Return type:

Figure