Skip to content

perturbation_space: skip None-keyed layers in _combine #1000

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Zethson merged 1 commit into from
May 26, 2026
Merged

perturbation_space: skip None-keyed layers in _combine #1000

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
27 changes: 19 additions & 8 deletions pertpy/tools/_perturbation_space/_perturbation_space.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -186,19 +186,28 @@ def _combine(

if ensure_consistency:
adata = self.compute_control_diff(adata, copy=True, all_data=True, target_col=target_col)
rename_back = {
key: key.removesuffix("_control_diff")
for key in [*adata.layers.keys(), *adata.obsm.keys()]
if key.endswith("_control_diff")
}
else:
warnings.warn(
"Combining perturbations without `ensure_consistency=True` is only well-defined "
"if the input was already differenced against control "
"(otherwise perturbation - perturbation != control).",
stacklevel=3,
)
rename_back = {}

# sc.get.aggregate can leave a `None`-keyed layer behind (the pre-aggregation .X);
# PseudobulkSpace.compute strips it but defensively re-strip here so callers passing
# a hand-built AnnData don't crash inside the string ops below.
layer_keys = [k for k in adata.layers if isinstance(k, str)]
obsm_keys = [k for k in adata.obsm if isinstance(k, str)]
rename_back = (
{
key: key.removesuffix("_control_diff")
for key in [*layer_keys, *obsm_keys]
if key.endswith("_control_diff")
}
if ensure_consistency
else {}
)

def _running(values: np.ndarray) -> np.ndarray:
result = values[adata.obs_names.get_loc(reference_key)].astype(float, copy=True)
Expand All @@ -207,13 +216,15 @@ def _running(values: np.ndarray) -> np.ndarray:
return result

new_layers: dict[str, np.ndarray] = {}
for layer_key, mat in adata.layers.items():
for layer_key in layer_keys:
mat = adata.layers[layer_key]
new_layers[rename_back.get(layer_key, layer_key)] = np.concatenate(
(np.asarray(mat), _running(np.asarray(mat))[None, :]), axis=0
)

new_obsm: dict[str, np.ndarray] = {}
for embedding_key, mat in adata.obsm.items():
for embedding_key in obsm_keys:
mat = adata.obsm[embedding_key]
new_obsm[rename_back.get(embedding_key, embedding_key)] = np.concatenate(
(np.asarray(mat), _running(np.asarray(mat))[None, :]), axis=0
)
Expand Down
Loading