ci: download small EBI dataset, use forksafe threadpool for docs (#4514)

Author: flying-sheep

src/scanpy/datasets/_ebi_expression_atlas.py

@@ -139,9 +139,9 @@ def ebi_expression_atlas(
     Example
     -------
     >>> import scanpy as sc
-    >>> sc.datasets.ebi_expression_atlas("E-MTAB-4888")  # doctest: +ELLIPSIS
-    AnnData object with n_obs × n_vars = 2261 × 23899
-        obs: 'Sample Characteristic[organism]', 'Sample Characteristic Ontology Term[organism]', ..., 'Factor Value[cell type]', 'Factor Value Ontology Term[cell type]'
+    >>> sc.datasets.ebi_expression_atlas("E-ENAD-19")  # doctest: +ELLIPSIS
+    AnnData object with n_obs × n_vars = 26 × 3644
+        obs: 'Sample Characteristic[organism]', 'Sample Characteristic Ontology Term[organism]', ..., 'Factor Value[single cell identifier]', 'Factor Value Ontology Term[single cell identifier]'
         layers: None (.X)
 
     """

tests/test_datasets.py

@@ -84,14 +84,15 @@ def test_ebi_expression_atlas(monkeypatch: pytest.MonkeyPatch):
     from scanpy.datasets import _ebi_expression_atlas as ea_mod
 
     # make sure we use chunks when testing.
-    # This dataset has <8M entries, so 4M entries/chunk = 2 chunks
+    # This dataset has ~31k entries, so 20k entries/chunk = 2 chunks
     assert hasattr(ea_mod, "CHUNK_SIZE")
-    monkeypatch.setattr(ea_mod, "CHUNK_SIZE", int(4e6))
+    monkeypatch.setattr(ea_mod, "CHUNK_SIZE", 20_000)
 
-    adata = sc.datasets.ebi_expression_atlas("E-MTAB-4888")
+    # a tiny dataset (~160kB), since EBI’s downloads can be very slow
+    adata = sc.datasets.ebi_expression_atlas("E-ENAD-19")
     # The shape changes sometimes
-    assert 2261 <= adata.shape[0] <= 2315
-    assert 23899 <= adata.shape[1] <= 24051
+    assert 24 <= adata.shape[0] <= 28
+    assert 3500 <= adata.shape[1] <= 3800
 
 
 @pytest.mark.parametrize(