ecotyper.rcompat.cluster¶

Hierarchical clustering and silhouettes exactly as in R.

Line-by-line ports of R’s stats Fortran routines HCLUST/HCASS2 (src/library/stats/src/hclust.f), cutree (hclust-utils.c), cophenetic.default, dist(method = "euclidean") (distance.c) and cluster::silhouette.default (sildist.c). The same loop order and strict comparisons are kept so that ties are broken identically.

Functions

`cophenetic`(tree)	`cophenetic(tree)` as a dist vector (as.dist order).
`cutree`(tree, k)	`cutree(tree, k)` → 1-based cluster ids in observation order.
`dist_euclidean_rows`(X)	`dist(X)` (Euclidean distances between rows of `X`).
`hclust`(d, n[, method, labels])	`hclust(d, method)` where `d` is a `dist` vector (as.dist order).
`silhouette_avg_width`(clustering, dmatrix)	`summary(silhouette(x, dmatrix = D))$avg.width` (NaN when R returns NA).
`silhouette_avg_width_dist`(clustering, d)	`summary(silhouette(x, d))$avg.width` for a `dist` vector.
`silhouette_widths`(clustering, d, ismat)	`cluster::silhouette.default`: widths in observation order, or None (NA).

Classes

HClust(merge, height, order[, labels, method])

class ecotyper.rcompat.cluster.HClust(merge: 'np.ndarray', height: 'np.ndarray', order: 'np.ndarray', labels: 'list | None' = None, method: 'str' = 'complete')[source]¶

Bases: object

Parameters:

merge (ndarray)
height (ndarray)
order (ndarray)
labels (list | None)
method (str)

merge: ndarray¶

height: ndarray¶

order: ndarray¶

labels: list | None = None¶

method: str = 'complete'¶

ecotyper.rcompat.cluster.hclust(d, n, method='complete', labels=None)[source]¶

hclust(d, method) where d is a dist vector (as.dist order).

Parameters:

d (ndarray)
n (int)
method (str)

Return type:

ecotyper.rcompat.cluster.cutree(tree, k)[source]¶

cutree(tree, k) → 1-based cluster ids in observation order.

Parameters:

tree (HClust)
k (int)

Return type:

ecotyper.rcompat.cluster.cophenetic(tree)[source]¶

cophenetic(tree) as a dist vector (as.dist order).

Parameters:: tree (HClust)
Return type:: ndarray

ecotyper.rcompat.cluster.dist_euclidean_rows(X)[source]¶

dist(X) (Euclidean distances between rows of X).

Parameters:: X (ndarray)
Return type:: ndarray

ecotyper.rcompat.cluster.silhouette_widths(clustering, d, ismat)[source]¶

cluster::silhouette.default: widths in observation order, or None (NA).

Parameters:: ismat (bool)
Return type:: ndarray | None

ecotyper.rcompat.cluster.silhouette_avg_width(clustering, dmatrix)[source]¶

summary(silhouette(x, dmatrix = D))$avg.width (NaN when R returns NA).

Parameters:: dmatrix (ndarray)
Return type:: float

ecotyper.rcompat.cluster.silhouette_avg_width_dist(clustering, d)[source]¶

summary(silhouette(x, d))$avg.width for a dist vector.

Parameters:: d (ndarray)
Return type:: float