tiledbsoma.io.from_anndata¶

tiledbsoma.io.from_anndata(experiment_uri: str, anndata: ~anndata._core.anndata.AnnData, measurement_name: str, *, context: ~tiledbsoma.options._soma_tiledb_context.SOMATileDBContext | None = None, platform_config: ~typing.Dict[str, ~typing.Mapping[str, ~typing.Any]] | object | None = None, obs_id_name: str = 'obs_id', var_id_name: str = 'var_id', X_layer_name: str = 'data', raw_X_layer_name: str = 'data', ingest_mode: ~typing.Literal['write', 'schema_only', 'resume'] = 'write', use_relative_uri: bool | None = None, X_kind: type[~tiledbsoma._sparse_nd_array.SparseNDArray] | type[~tiledbsoma._dense_nd_array.DenseNDArray] = <class 'tiledbsoma._sparse_nd_array.SparseNDArray'>, registration_mapping: ~tiledbsoma.io._registration.ambient_label_mappings.ExperimentAmbientLabelMapping | None = None, uns_keys: ~collections.abc.Sequence[str] | None = None, additional_metadata: dict[str, bytes | float | int | str] | None = None) → str¶

Writes an AnnData object to an Experiment.

This function is similar to from_h5ad() except that it operates on an AnnData object already loaded into memory, rather than reading from an .h5ad file on disk.

Measurement data is stored in a Measurement in the experiment’s ms field, with the key provided by measurement_name. Data elements are available at the standard fields (var, X, etc.). Unstructured data from uns is partially supported (structured arrays and non-numeric NDArrays are skipped), and is available at the measurement’s uns key (i.e., at your_experiment.ms[measurement_name]["uns"]).

Parameters:

experiment_uri – The experiment to create or update.
anndata – An AnnData object already loaded into memory.
measurement_name – The name of the measurement to store data in.
context – Optional SOMATileDBContext containing storage parameters, etc.
platform_config – Platform-specific options used to create this array. See from_h5ad() for details.
obs_id_name/var_id_name – Column names to use for obs/var IDs in append mode. See from_h5ad() for detailed behavior.
X_layer_name – SOMA array name for the AnnData’s X matrix.
raw_X_layer_name – SOMA array name for the AnnData’s raw/X matrix.
ingest_mode – One of "write", "resume" (deprecated), or "schema_only". See from_h5ad() for details.
use_relative_uri – Whether to use relative URIs for nested objects.
X_kind – Matrix type for storing X data: DenseNDArray or SparseNDArray.
registration_mapping – Pre-computed registration for multi-dataset ingestion. For multi-dataset workflows, use register_anndatas() to create this mapping. See from_h5ad() for usage examples.
uns_keys – Only ingest the specified top-level uns keys. Default is to ingest all. Use uns_keys=[] to skip uns entirely.
additional_metadata – Optional metadata to add to the Experiment and all descendents. See from_h5ad() for details and examples.

Returns:

The URI of the newly created experiment.