relevanceai.operations.auto.dr#

Module Contents#

class relevanceai.operations.auto.dr.DimensionalityReduction(*args, **kw)#

A Pandas Like datatset API for interacting with the RelevanceAI python package

auto_reduce_dimensions(self, alias: str, vector_fields: list, filters: Optional[list] = None, number_of_documents: Optional[int] = None)#

Run dimensionality reduction quickly on a dataset on a small number of documents. This is useful if you want to quickly see a projection of your dataset.

Warning

This function is currently in beta and is likely to change in the future. We recommend not using this in any production systems.

Parameters
  • vector_fields (list) – The vector fields to run dimensionality reduction on

  • number_of_documents (int) – The number of documents to get

  • algorithm (str) – The algorithm to run. The only supported algorithm is pca at this current point in time.

  • n_components (int) – The number of components

Example

from relevanceai import Client
client = Client()
df = client.Dataset("sample")
df.auto_reduce_dimensions(
    "pca-3",
    ["sample_vector_"],
)
reduce_dimensions(self, vector_fields: list, alias: str, number_of_documents: int = 1000, algorithm: str = 'pca', n_components: int = 3, filters: Optional[list] = None)#

Run dimensionality reduction quickly on a dataset on a small number of documents. This is useful if you want to quickly see a projection of your dataset.

Warning

This function is currently in beta and is likely to change in the future. We recommend not using this in any production systems.

Parameters
  • vector_fields (list) – The vector fields to run dimensionality reduction on

  • number_of_documents (int) – The number of documents to get

  • algorithm (str) – The algorithm to run. The only supported algorithm is pca at this current point in time.

  • n_components (int) – The number of components

Example

from relevanceai import Client
client = Client()
df = client.Dataset("sample")
df.auto_reduce_dimensions(
    alias="pca-3",
    ["sample_vector_"],
    number_of_documents=1000
)