marsilea.base.ClusterBoard#

class marsilea.base.ClusterBoard(cluster_data, width=None, height=None, name=None, margin=0.2)#

Bases: WhiteBoard

add_dendrogram(side, method=None, metric=None, linkage=None, add_meta=True, add_base=True, add_divider=True, meta_color=None, linewidth=None, colors=None, divider_style='--', meta_ratio=0.2, show=True, name=None, size=0.5, pad=0.0, get_meta_center=None)#

Run cluster and add dendrogram

Note

method and metric only works when you add the first row/col dendrogram.
If add_meta=False and add_base=False, the dendrogram axes will not be created.

Parameters:

side
methodstr: See scipy’s linkage
metricstr: See scipy’s linkage
linkagendarray: Precomputed linkage matrix. See scipy’s linkage for specific format.
add_metaNone | bool: By default, add_meta is set to False if the linkage is provided, otherwise True. If the data is split, a meta dendrogram can be drawn for data chunks. The mean value of the data chunk is used to calculate linkage matrix for meta dendrogram.
add_baseNone | bool: By default, add_meta is set to False if the linkage is provided, otherwise True. Draw the base dendrogram for each data chunk. You can turn this off if the base dendrogram is too crowded.
add_dividerbool: Draw a divide line that tells the difference between base and meta dendrogram.
divider_stylestr, default: “–”: The line style of the divide line
meta_colorcolor: The color of the meta dendrogram
meta_ratiofloat: The size of meta dendrogram relative to the base dendrogram
linewidthfloat: The linewidth for every dendrogram and divide line
colorscolor, array of color: If only one color is specified, it will be applied to every dendrogram. If an array of color is specified, it will be applied to each base dendrogram.
showbool: If False, the dendrogram will not be drawn and the axes will not be created.
namestr: The name of the dendrogram axes
sizefloat
padfloat
get_meta_center: callable: A function to calculate the centroid of data. It should accept a 2D numpy array as input and return a 1D numpy array of the same length as the number of columns in the input, representing the centroid. The default will use the mean values.

Examples

You can change how the linkage matrix is calculated

>>> data = np.random.rand(10, 11)
>>> import marsilea as ma
>>> h = ma.Heatmap(data)
>>> h.add_dendrogram("left", method="ward", colors="green")
>>> h.render()

../../_images/marsilea-base-ClusterBoard-1.png

Only show the meta dendrogram to avoid crowded dendrogram

>>> h = ma.Heatmap(data)
>>> h.hsplit(cut=[4, 8])
>>> h.add_dendrogram("left", add_base=False)
>>> h.render()

../../_images/marsilea-base-ClusterBoard-2.png

Change color for each base dendrogram

>>> h = ma.Heatmap(data)
>>> h.hsplit(cut=[4, 8])
>>> h.add_dendrogram("left", colors=["#5470c6", "#91cc75", "#fac858"])
>>> h.render()

../../_images/marsilea-base-ClusterBoard-3.png

property col_cluster#

get_col_linkage()#

get_deform()#

get_row_linkage()#

hsplit(cut=None, labels=None, order=None, spacing=0.01)#

property row_cluster#

square = False#

vsplit(cut=None, labels=None, order=None, spacing=0.01)#