sfaira.data.DatasetGroup

class sfaira.data.DatasetGroup(datasets: dict, collection_id: str = 'default')

Container class that co-manages multiple data sets, removing need to call Dataset() methods directly through wrapping them.

Example:

#query loaders lung #from sfaira.dev.data.loaders.lung import DatasetGroupLung as DatasetGroup #dsg_humanlung = DatasetGroupHuman(path=’path/to/data’) #dsg_humanlung.load_all(match_to_reference=’Homo_sapiens_GRCh38_97’) #dsg_humanlung[some_id] #dsg_humanlung.adata

Attributes

adata
adata_ls
additional_annotation_key	"
collection_id
doi	Propagates DOI annotation from contained datasets.
ids
ontology_celltypes	use might be replaced by ontology_container_sfaira in the future.
ontology_container_sfaira
supplier	Propagates supplier annotation from contained datasets.
datasets

Methods

collapse_counts()	Collapse count matrix along duplicated index.
download(**kwargs)
load([annotated_only, load_raw, ...])	Load all datasets in group (option for temporary loading).
ncells([annotated_only])
ncells_bydataset([annotated_only])
obs_concat([keys])	Returns concatenation of all .obs.
project_celltypes_to_ontology([...])	Project free text cell type names to ontology based on mapping table.
show_summary()
streamline_features([match_to_release, ...])	Subset and sort genes to genes defined in an assembly or genes of a particular type, such as protein coding. :param match_to_release: Which genome annotation release to map the feature space to. Note that assemblies from ensbeml are usually named as Organism.Assembly.Release, this is the Release string. Can be: - str: Provide the name of the release. - dict: Mapping of organism to name of the release (see str format). Chooses release for each data set based on organism annotation.:param remove_gene_version: Whether to remove the version number after the colon sometimes found in ensembl gene ids. :param subset_genes_to_type: Type(s) to subset to. Can be a single type or a list of types or None. Types can be: - None: All genes in assembly. - "protein_coding": All protein coding genes in assembly.
streamline_metadata([schema, clean_obs, ...])	Streamline the adata instance in each data set to output format.
subset(key, values)	Subset list of adata objects based on sample-wise properties.
subset_cells(key, values)	Subset list of adata objects based on cell-wise properties.
write_backed(adata_backed, genome, idx[, ...])	Loads data set group into slice of backed anndata object.
write_distributed_store(dir_cache[, ...])	Write data set into a format that allows distributed access to data set on disk.
write_ontology_class_maps(fn, attrs[, ...])	Write cell type maps of free text cell types to ontology classes.