Merge branch 'main' into pre-commit-ci-update-config

docs/references.bib

@@ -440,3 +440,14 @@ @article{klein:23
   title     = {Mapping cells through time and space with moscot},
   year      = {2023},
 }
+
+@article{srivatsan:20,
+  title={Massively multiplex chemical transcriptomics at single-cell resolution},
+  author={Srivatsan, Sanjay R and McFaline-Figueroa, Jos{\'e} L and Ramani, Vijay and Saunders, Lauren and Cao, Junyue and Packer, Jonathan and Pliner, Hannah A and Jackson, Dana L and Daza, Riza M and Christiansen, Lena and others},
+  journal={Science},
+  volume={367},
+  number={6473},
+  pages={45--51},
+  year={2020},
+  publisher={American Association for the Advancement of Science}
+}

docs/user.rst

@@ -51,6 +51,7 @@ Datasets
     datasets.drosophila
     datasets.hspc
     datasets.mosta
+    datasets.sciplex
     datasets.sim_align
     datasets.simulate_data
     datasets.tedsim

src/moscot/datasets.py

@@ -307,6 +307,38 @@ def tedsim(
     )
 
 
+def sciplex(
+    path: PathLike = "~/.cache/moscot/sciplex.h5ad",
+    force_download: bool = False,
+    **kwargs: Any,
+) -> AnnData:  # pragma: no cover
+    """Perturbation dataset published in :cite:`srivatsan:20`.
+
+    Transcriptomes of A549, K562, and mCF7 cells exposed to 188 compounds.
+    Data obtained from http://projects.sanderlab.org/scperturb/.
+
+    Parameters
+    ----------
+    path
+        Path where to save the file.
+    force_download
+        Whether to force-download the data.
+    kwargs
+        Keyword arguments for :func:`scanpy.read`.
+
+    Returns
+    -------
+    Annotated data object.
+    """
+    return _load_dataset_from_url(
+        path,
+        backup_url="https://figshare.com/ndownloader/files/43381398",
+        expected_shape=(799317, 110984),
+        force_download=force_download,
+        **kwargs,
+    )
+
+
 def sim_align(
     path: PathLike = "~/.cache/moscot/sim_align.h5ad",
     force_download: bool = False,