Backport PR #3351: Simplify scale

docs/release-notes/3351.bugfix.md

@@ -0,0 +1 @@
+Fix zappy compatibility for clip_array {smaller}`P Angerer`

src/scanpy/_utils/__init__.py

@@ -62,6 +62,8 @@
     _MemoryArray = NDArray | CSBase
     _SupportedArray = _MemoryArray | DaskArray
 
+    _SA = TypeVar("_SA", bound=_SupportedArray)
+
     _ForT = TypeVar("_ForT", bound=Callable | type)
 
 
@@ -273,9 +275,7 @@ def get_igraph_from_adjacency(adjacency: CSBase, *, directed: bool = False) -> G
     import igraph as ig
 
     sources, targets = adjacency.nonzero()
-    weights = adjacency[sources, targets]
-    if isinstance(weights, np.matrix):
-        weights = weights.A1
+    weights = dematrix(adjacency[sources, targets]).ravel()
     g = ig.Graph(directed=directed)
     g.add_vertices(adjacency.shape[0])  # this adds adjacency.shape[0] vertices
     g.add_edges(list(zip(sources, targets, strict=True)))
@@ -836,6 +836,14 @@ def _check_nonnegative_integers_dask(X: DaskArray) -> DaskArray:
     return X.map_blocks(check_nonnegative_integers, dtype=bool, drop_axis=(0, 1))
 
 
+def dematrix(x: _SA | np.matrix) -> _SA:
+    if isinstance(x, np.matrix):
+        return x.A
+    if isinstance(x, DaskArray) and isinstance(x._meta, np.matrix):
+        return x.map_blocks(np.asarray, meta=np.array([], dtype=x.dtype))
+    return x
+
+
 def select_groups(
     adata: AnnData,
     groups_order_subset: Iterable[str] | Literal["all"] = "all",

src/scanpy/preprocessing/_deprecated/__init__.py

@@ -4,6 +4,7 @@
 from scipy import sparse
 
 from ..._compat import CSBase, old_positionals
+from ..._utils import dematrix
 
 
 @old_positionals("max_fraction", "mult_with_mean")
@@ -39,15 +40,11 @@ def normalize_per_cell_weinreb16_deprecated(
         msg = "Choose max_fraction between 0 and 1."
         raise ValueError(msg)
 
-    counts_per_cell = x.sum(1)
-    if isinstance(counts_per_cell, np.matrix):
-        counts_per_cell = counts_per_cell.A1
-    gene_subset = np.all(x <= counts_per_cell[:, None] * max_fraction, axis=0)
-    if isinstance(gene_subset, np.matrix):
-        gene_subset = gene_subset.A1
-    tc_include = x[:, gene_subset].sum(1)
-    if isinstance(tc_include, np.matrix):
-        tc_include = tc_include.A1
+    counts_per_cell = dematrix(x.sum(1)).ravel()
+    gene_subset = dematrix(
+        np.all(x <= counts_per_cell[:, None] * max_fraction, axis=0)
+    ).ravel()
+    tc_include = dematrix(x[:, gene_subset].sum(1)).ravel()
 
     x_norm = (
         x.multiply(sparse.csr_matrix(1 / tc_include[:, None]))  # noqa: TID251

src/scanpy/preprocessing/_normalization.py

@@ -8,7 +8,7 @@
 
 from .. import logging as logg
 from .._compat import CSBase, DaskArray, old_positionals
-from .._utils import axis_mul_or_truediv, axis_sum, view_to_actual
+from .._utils import axis_mul_or_truediv, axis_sum, dematrix, view_to_actual
 from ..get import _get_obs_rep, _set_obs_rep
 
 try:
@@ -216,9 +216,8 @@ def normalize_total(  # noqa: PLR0912, PLR0915
         counts_per_cell = np.ravel(counts_per_cell)
 
         # at least one cell as more than max_fraction of counts per cell
-
-        gene_subset = axis_sum((x > counts_per_cell[:, None] * max_fraction), axis=0)
-        gene_subset = np.asarray(np.ravel(gene_subset) == 0)
+        hi_exp = dematrix(x > counts_per_cell[:, None] * max_fraction)
+        gene_subset = axis_sum(hi_exp, axis=0) == 0
 
         msg += (
             ". The following highly-expressed genes are not considered during "