Skip to content

output

Transforms applied to simulated counts after sampling.

ObservedMaxCap

Caps features at the maximum observed during training

Attributes:

Name Type Description
maximum ndarray of shape (n_features,) or None

Per-feature maximum from the last successful :meth:fit; None before the first.

Source code in scdesigner/src/scdesigner/simulators/output.py
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
class ObservedMaxCap:
    """Caps features at the maximum observed during training

    Attributes
    ----------
    maximum : ndarray of shape (n_features,) or None
        Per-feature maximum from the last successful :meth:`fit`; ``None``
        before the first.
    """

    def __init__(self):
        self.maximum: Optional[np.ndarray] = None

    def fit(self, adata: AnnData, *, chunk_size: Optional[int] = None) -> None:
        """Store the per-feature maximum of ``adata.X``.

        Parameters
        ----------
        adata : AnnData
            In memory or backed.
        chunk_size : int, optional
            See :func:`~scdesigner.data.feature_summary.summarize_features`.
        """
        self.maximum = summarize_features(adata, include_max=True, chunk_size=chunk_size).maximum

    def transform(self, samples: np.ndarray) -> np.ndarray:
        """New array of ``samples`` capped at ``maximum``.

        Parameters
        ----------
        samples : ndarray of shape (n_obs, n_features)
        """
        if self.maximum is None:
            raise RuntimeError("ObservedMaxCap must be fit before transform.")
        return np.minimum(samples, self.maximum)

fit(adata, *, chunk_size=None)

Store the per-feature maximum of adata.X.

Parameters:

Name Type Description Default
adata AnnData

In memory or backed.

required
chunk_size int

See :func:~scdesigner.data.feature_summary.summarize_features.

None
Source code in scdesigner/src/scdesigner/simulators/output.py
22
23
24
25
26
27
28
29
30
31
32
def fit(self, adata: AnnData, *, chunk_size: Optional[int] = None) -> None:
    """Store the per-feature maximum of ``adata.X``.

    Parameters
    ----------
    adata : AnnData
        In memory or backed.
    chunk_size : int, optional
        See :func:`~scdesigner.data.feature_summary.summarize_features`.
    """
    self.maximum = summarize_features(adata, include_max=True, chunk_size=chunk_size).maximum

transform(samples)

New array of samples capped at maximum.

Parameters:

Name Type Description Default
samples ndarray of shape (n_obs, n_features)
required
Source code in scdesigner/src/scdesigner/simulators/output.py
34
35
36
37
38
39
40
41
42
43
def transform(self, samples: np.ndarray) -> np.ndarray:
    """New array of ``samples`` capped at ``maximum``.

    Parameters
    ----------
    samples : ndarray of shape (n_obs, n_features)
    """
    if self.maximum is None:
        raise RuntimeError("ObservedMaxCap must be fit before transform.")
    return np.minimum(samples, self.maximum)