Merge pull request #222 from scverse/add_bsize_psbulk

Add bsize psbulk

Author: PauBadiaM

src/decoupler/mt/_gsea.py

@@ -110,8 +110,8 @@ def _nesrank(
         neg_null_mean = null[neg_null_msk].mean()
         nes = -es / neg_null_mean
     else:
-        nes = np.inf
-        pval = np.inf
+        nes = 0.0
+        pval = 1.0
     return nes, pval
 
 

src/decoupler/pl/_filter_samples.py

@@ -12,7 +12,7 @@
 @docs.dedent
 def filter_samples(
     adata: AnnData,
-    groupby: str | list,
+    groupby: str | list | None = None,
     log: bool = True,
     min_cells: int | float = 10,
     min_counts: int | float = 1000,
@@ -51,10 +51,11 @@ def filter_samples(
     assert all(col in adata.obs.columns for col in ["psbulk_cells", "psbulk_counts"]), (
         "psbulk_* columns not present in adata.obs, this function should be used after running decoupler.pp.pseudobulk"
     )
-    assert isinstance(groupby, str | list), "groupby must be str or list"
+    assert isinstance(groupby, str | list) or groupby is None, "groupby must be str, list or None"
     if isinstance(groupby, str):
         groupby = [groupby]
-    assert all(col in adata.obs for col in groupby), "columns in groupby must be in adata.obs"
+    if groupby:
+        assert all(col in adata.obs for col in groupby), "columns in groupby must be in adata.obs"
     # Extract obs
     df = adata.obs.copy()
     # Transform to log10
@@ -65,7 +66,7 @@ def filter_samples(
         label_x, label_y = r"$\log_{10}$ " + label_x, r"$\log_{10}$ " + label_y
         min_cells, min_counts = np.log10(min_cells), np.log10(min_counts)
     # Plot
-    if len(groupby) > 1:
+    if groupby is not None and (len(groupby) > 1):
         # Instance
         assert kwargs.get("ax") is None, "when groupby is list, ax must be None"
         kwargs["ax"] = None
@@ -85,12 +86,14 @@ def filter_samples(
             ax.axhline(y=min_counts, linestyle="--", color="black")
     else:
         # Instance
-        groupby = groupby[0]
+        if isinstance(groupby, list):
+            groupby = groupby[0]
         bp = Plotter(**kwargs)
         bp.ax.grid(zorder=0)
         bp.ax.set_axisbelow(True)
         sns.scatterplot(x="psbulk_cells", y="psbulk_counts", hue=groupby, ax=bp.ax, data=df, zorder=1)
-        bp.ax.legend(loc="center left", bbox_to_anchor=(1, 0.5), frameon=False, title=groupby)
+        if groupby:
+            bp.ax.legend(loc="center left", bbox_to_anchor=(1, 0.5), frameon=False, title=groupby)
         bp.ax.set_xlabel(label_x)
         bp.ax.set_ylabel(label_y)
         bp.ax.axvline(x=min_cells, linestyle="--", color="black")

src/decoupler/pp/anndata.py

@@ -311,6 +311,7 @@ def pseudobulk(
     empty: bool = False,
     mode: str | Callable | dict = "sum",
     skip_checks: bool = False,
+    bsize: int = 250_000,
     verbose: bool = False,
 ) -> AnnData:
     """
@@ -365,7 +366,10 @@ def pseudobulk(
     assert isinstance(groups_col, str | list) or groups_col is None, "groups_col must be str or None"
     assert isinstance(mode, str | dict) or callable(mode), "mode must be str, dict or callable"
     # Extract data
-    X, obs, var = extract(adata, layer=layer, raw=raw, empty=empty, verbose=verbose)
+    X, obs, var = extract(adata, layer=layer, raw=raw, empty=empty, bsize=bsize, verbose=verbose)
+    assert len(set(obs)) == len(obs), (
+        "Repeated elements in adata.obs_names, to make them unique run adata.obs_names_make_unique()"
+    )
     obs = adata.obs.loc[obs].copy()
     var = adata.var.loc[var]
     # Validate X