hubmapconsortium / hubmap-api-py-client

Python client for the Cells API
MIT License
1 stars 0 forks source link
gehlenborglab hidivelab

hubmap-api-py-client

PyPI version

Python client for the HuBMAP Cells API; See also: hubmap-api-js-client.

Contributors start here.

Usage

Install from pypi:

pip install hubmap-api-py-client

Find cells with different criteria, and intersect resulting sets:

>>> from os import environ
>>> from hubmap_api_py_client import Client
>>> client = Client(environ['API_ENDPOINT'])

>>> [m for m in dir(client) if m.startswith('select_')]
['select_cells', 'select_celltypes', 'select_clusters', 'select_datasets', 'select_genes', 'select_organs', 'select_proteins']

>>> gene_symbol = client.select_genes(where="modality", has=["rna"]).get_list()[0]['gene_symbol']
>>> cells_with_gene = client.select_cells(where='gene', has=[f'{gene_symbol} > 0.5'], genomic_modality='rna')
>>> assert len(cells_with_gene) > 0

# Select cells from the datasets with the following UUIDs:
>>> dataset_a_uuid = client.select_datasets(where="gene", has=[f'{gene_symbol} > 1'], genomic_modality="rna", min_cell_percentage=0.0).get_list()[0]['uuid']
>>> dataset_b_uuid = client.select_datasets(where="gene", has=[f'{gene_symbol} > 1'], genomic_modality="rna", min_cell_percentage=0.0).get_list()[1]['uuid']
>>> cells_in_a_len = len(client.select_cells(where='dataset', has=[dataset_a_uuid]))
>>> cells_in_b_len = len(client.select_cells(where='dataset', has=[dataset_b_uuid]))
>>> cells_in_datasets = client.select_cells(where='dataset', has=[dataset_a_uuid, dataset_b_uuid])
>>> cells_in_datasets_len = len(cells_in_datasets)
>>> assert cells_in_datasets_len > 0
>>> assert cells_in_datasets_len == cells_in_a_len + cells_in_b_len

# Combine criteria with intersection:
>>> cells_with_gene_in_datasets = cells_with_gene & cells_in_datasets

# Get a list; should run quickly:
>>> cell_list = cells_with_gene.get_list()

>>> cells = cell_list[0:10]
>>> assert len(cells) > 0
>>> assert cells[0].keys() == {'cell_id', 'modality', 'dataset', 'organ', 'cell_type' ,'clusters'}

More documentation:

Only some types of objects can be retrieved from other types of objects:

where=... None cell cluster dataset gene organ protein modality celltype
select_cells()
select_clusters() ✓ ✩
select_datasets()
select_genes() ✓ ✩ ✓ ✩
select_organs() ✓ ✩
select_proteins()
select_celltypes()