`relevanceai.visualise.projector`

Module Contents

Classes

Projector

Projector class.

Attributes

`chained_assignment`
`RELEVANCEAI_BLUE`

relevanceai.visualise.projector.chained_assignment

relevanceai.visualise.projector.RELEVANCEAI_BLUE = #1854FF

class relevanceai.visualise.projector.Projector(project, api_key)

Bases: relevanceai.utils.Utils, relevanceai.api.client.BatchAPIClient, relevanceai.base._Base, doc_utils.DocUtils

Projector class.

Example

>>> from relevanceai import Client
>>> project = input()
>>> api_key = input()
>>> client = Client(project, api_key)
>>> client.projector.plot(
        dataset_id, vector_field, number_of_points_to_render, random_state,
        dr, dr_args, dims,
        vector_label, label_char_length,
        color_label, colour_label_char_length,
        hover_label,
        cluster, cluster_args,
        )

plot(self, dataset_id: str, vector_field: str, number_of_points_to_render: int = 1000, vector_label: relevanceai.vector_tools.constants.Union[None, str] = None, dr: relevanceai.vector_tools.constants.Union[relevanceai.vector_tools.constants.DIM_REDUCTION, relevanceai.vector_tools.dim_reduction.DimReductionBase] = 'pca', dims: relevanceai.vector_tools.constants.Literal[2, 3] = 3, dr_args: relevanceai.vector_tools.constants.Union[None, relevanceai.vector_tools.constants.Dict] = None, cluster: relevanceai.vector_tools.constants.Union[relevanceai.vector_tools.constants.CLUSTER, relevanceai.vector_tools.cluster.ClusterBase] = None, num_clusters: relevanceai.vector_tools.constants.Union[None, int] = 10, cluster_args: relevanceai.vector_tools.constants.Dict = {}, cluster_on_dr: bool = False, hover_label: list = [], show_image: bool = False, label_char_length: int = 50, marker_size: int = 5, interactive: bool = False)

Dimension reduce vectors and plot with functionality to visualise different clusters and nearest neighbours

To write your own custom dimensionality reduction, you should inherit from DimReductionBase: from relevanceai.visualise.dim_reduction import DimReductionBase class CustomDimReduction(DimReductionBase):

def fit_transform(self, vectors):
return np.arange(512, 2)

Example

>>> from relevanceai import Client
>>> project = input()
>>> api_key = input()
>>> client = Client(project, api_key)
>>> client.projector.plot(
        dataset_id, vector_field, number_of_points_to_render, random_state,
        dr, dr_args, dims,
        vector_label, label_char_length,
        color_label, colour_label_char_length,
        hover_label,
        cluster, cluster_args,
        )

Parameters

dataset_id (string) – Unique name of dataset
vector_field (list) – Vector field to plot
number_of_points_to_render (int) – Number of vector fields to plot
vector_label (string) – Field to use as label to describe vector on plot
dr (string) – Method of dimension reduction for vectors
dims (int) – Number of dimensions to reduce to
dr_args (dict) – Additional arguments for dimension reduction
cluster (string) – Method of clustering for vectors
num_clusters (string) – Number of clusters to create
cluster_args (dict) – Additional arguments for clustering
cluster_on_dr (int) – Whether to cluster on the dimension reduced or original vectors
hover_label (list) – Additional labels to include as plot labels
show_image (bool) – Whether vector labels are image urls
label_char_length (int) – Maximum length of text for each hover label
marker_size (int) – Marker size of the plot
interactive (bool) – Whether to include interactive features including nearest neighbours

plot_from_docs(self, docs: relevanceai.vector_tools.constants.List[relevanceai.vector_tools.constants.Dict], vector_field: str, vector_label: relevanceai.vector_tools.constants.Union[None, str] = None, dr: relevanceai.vector_tools.constants.Union[relevanceai.vector_tools.constants.DIM_REDUCTION, relevanceai.vector_tools.dim_reduction.DimReductionBase] = 'pca', dims: relevanceai.vector_tools.constants.Literal[2, 3] = 3, dr_args: relevanceai.vector_tools.constants.Union[None, relevanceai.vector_tools.constants.Dict] = None, cluster: relevanceai.vector_tools.constants.Union[relevanceai.vector_tools.constants.CLUSTER, relevanceai.vector_tools.cluster.ClusterBase] = None, num_clusters: relevanceai.vector_tools.constants.Union[None, int] = 10, cluster_args: relevanceai.vector_tools.constants.Dict = {}, cluster_on_dr: bool = False, hover_label: list = [], show_image: bool = False, label_char_length: int = 50, marker_size: int = 5, dataset_name: relevanceai.vector_tools.constants.Union[None, str] = None, interactive: bool = False)

_generate_plot_data(self, embedding_df: pandas.DataFrame, hover_label: relevanceai.vector_tools.constants.List[str], dims: int, marker_size: int, cluster: relevanceai.vector_tools.constants.Union[relevanceai.vector_tools.constants.Literal[kmeans], relevanceai.vector_tools.constants.Literal[kmedoids], relevanceai.vector_tools.constants.Literal[hdbscan], relevanceai.vector_tools.cluster.ClusterBase, None], label_char_length: int)

_generate_layout(self, plot_title)

_generate_plot_info(self, embedding_df, hover_label, dims, marker_size, label_char_length)

_generate_hover_template(self, df: pandas.DataFrame, dims: int, hover_label: list, label_char_length: int): Generating hover template

_generate_plot_title(self, dims, dataset_name, number_of_points, vector_field, vector_label, label_char_length)

relevanceai.visualise.projector

Module Contents

Classes

Attributes

`relevanceai.visualise.projector`