Augur scsim warnings

Signed-off-by: zethson <[email protected]>

.gitignore

@@ -155,3 +155,5 @@ lightning_logs/*
 */lightning_logs/*
 
 node_modules
+
+test.ipynb

pertpy/data/_datasets.py

@@ -66,7 +66,7 @@ def sc_sim_augur() -> AnnData:  # pragma: no cover
     output_file_path = settings.datasetdir / output_file_name
     if not Path(output_file_path).exists():
         _download(
-            url="https://figshare.com/ndownloader/files/31645886",
+            url="https://figshare.com/ndownloader/files/49828902",
             output_file_name=output_file_name,
             output_path=settings.datasetdir,
             is_zip=False,

pertpy/tools/_augur.py

@@ -25,11 +25,11 @@
     explained_variance_score,
     f1_score,
     make_scorer,
-    mean_squared_error,
     precision_score,
     r2_score,
     recall_score,
     roc_auc_score,
+    root_mean_squared_error,
 )
 from sklearn.model_selection import StratifiedKFold, cross_validate
 from sklearn.preprocessing import LabelEncoder
@@ -440,7 +440,7 @@ def set_scorer(
                 "augur_score": make_scorer(self.ccc_score),
                 "r2": make_scorer(r2_score),
                 "ccc": make_scorer(self.ccc_score),
-                "neg_mean_squared_error": make_scorer(mean_squared_error),
+                "neg_mean_squared_error": make_scorer(root_mean_squared_error),
                 "explained_variance": make_scorer(explained_variance_score),
             }
         )

pyproject.toml

@@ -125,7 +125,8 @@ addopts = [
 ]
 markers = [
     "conda: marks a subset of tests to be ran on the Bioconda CI.",
-    "extra: marks tests that require extra dependencies."
+    "extra: marks tests that require extra dependencies.",
+    "slow: mark test as slow to run"
 ]
 minversion = 6.0
 norecursedirs = [ '.*', 'build', 'dist', '*.egg', 'data', '__pycache__']

tests/tools/test_augur.py

@@ -6,7 +6,6 @@
 import pytest
 import scanpy as sc
 from pertpy.tools._augur import Params
-from sklearn.ensemble import RandomForestRegressor
 
 CWD = Path(__file__).parent.resolve()
 
@@ -145,7 +144,7 @@ def test_select_variance(adata):
     """Test select variance implementation."""
     adata = ag_rfc.load(adata)
     sc.pp.highly_variable_genes(adata)
-    adata_cell_type = adata[adata.obs["cell_type"] == "CellTypeA"]
+    adata_cell_type = adata[adata.obs["cell_type"] == "CellTypeA"].copy()
     ad = ag_rfc.select_variance(adata_cell_type, var_quantile=0.5, span=0.3, filter_negative_residuals=False)
 
     assert 3672 == len(ad.var.index[ad.var["highly_variable"]])