tfds.visualization.GraphVisualizer

Visualizer for graph prediction datasets.

Inherits From: Visualizer

Methods

match

View source

match(
    ds_info: dataset_info.DatasetInfo
) -> bool

Checks whether this dataset can be visualized with this visualizer.

See base class for more information. Args: ds_info: Metadata for the dataset.

Returns
boolean value indicating whether the current visualizer can be used.

show

View source

show(
    ds: tf.data.Dataset,
    ds_info: dataset_info.DatasetInfo,
    node_label_fn: Optional[_GraphFn] = None,
    node_color_fn: Optional[_GraphFn] = None,
    edge_color_fn: Optional[_GraphFn] = None,
    rows: int = 3,
    cols: int = 3,
    plot_scale: float = 5,
    **kwargs
)

Display the dataset.

Args
ds tf.data.Dataset. The tf.data.Dataset object to visualize. Examples should not be batched. Examples will be consumed in order until (rows * cols) are read or the dataset is consumed.
ds_info tfds.core.DatasetInfo object of the dataset to visualize.
node_label_fn A callable that maps individual graph examples to a dictionary of node labels, rendered within the nodes.
node_color_fn A callable that maps individual graph examples to a dictionary of node colors.
edge_color_fn A callable that maps individual graph examples to a dictionary of edge colors.
rows int, number of rows of the display grid.
cols int, number of columns of the display grid.
plot_scale float, controls the plot size of the images. Keep this value around 5 to get a good plot. High and low values may cause the labels to get overlapped.
**kwargs Additional arguments passed to networkx.draw_networkx().

Returns
fig The pyplot figure.