weaviate.contextionary

Contextionary module used to interact with Weaviate’s contextionary module.

class weaviate.contextionary.Contextionary(connection: weaviate.connect.connection.Connection)

Bases: object

Contextionary class used to add extend the Weaviate contextionary module or to get vector/s of a specific concept.

Initialize a Contextionary class instance.

Parameters

connection (weaviate.connect.Connection) – Connection object to an active and running Weaviate instance.

extend(concept: str, definition: str, weight: float = 1.0) None

Extend the text2vec-contextionary with new concepts

Parameters
  • concept (str) – The new concept that should be added that is not in the Weaviate or needs to be updated, e.g. an abbreviation.

  • definition (str) – The definition of the new concept.

  • weight (float, optional) – The weight of the new definition compared to the old one, must be in-between the interval [0.0; 1.0], by default 1.0

Examples

>>> client.contextionary.extend(
...     concept = 'palantir',
...     definition = 'spherical stone objects used for communication in Middle-earth'
... )
Raises
  • TypeError – If an argument is not of an appropriate type.

  • ValueError – If ‘weight’ is outside the interval [0.0; 1.0].

  • requests.ConnectionError – If text2vec-contextionary could not be extended.

  • weaviate.UnexpectedStatusCodeException – If the network connection to weaviate fails.

get_concept_vector(concept: str) dict

Retrieves the vector representation of the given concept.

Parameters

concept (str) – Concept for which the vector should be retrieved. May be camelCase for word combinations.

Examples

>>> client.contextionary.get_concept_vector('king')
{
    "individualWords": [
        {
        "info": {
            "nearestNeighbors": [
            {
                "word": "king"
            },
            {
                "distance": 5.7498446,
                "word": "kings"
            },
            ...,
            {
                "distance": 6.1396513,
                "word": "queen"
            }
            ],
            "vector": [
            -0.68988,
            ...,
            -0.561865
            ]
        },
        "present": true,
        "word": "king"
        }
    ]
}
Returns

A dictionary containing info and the vector/s of the concept. The vector might be empty if the text2vec-contextionary does not contain it.

Return type

dict

Raises