bob_vector_db.qdrant_embedder

Classes

QdrantEmbedder

Qdrant JSON embedding interface.

Module Contents

class bob_vector_db.qdrant_embedder.QdrantEmbedder(model: str, **kwargs)

Bases: json_embedder.JsonEmbedder

Qdrant JSON embedding interface.

The __call__ function expects a string with a JSON dict containing the following attributes:: collection: to be used collection name. documents: list with document content, if also images are provided this should be the caption. metadatas: list of metadata dictionaries used as payload. images: optional list of image file names to embed along with the documents. ids: optional list of id strings, if not provided automatically uuid4 are created.

model

text_model = None

image_model = None

__call__(j: str, media: list = None) → None

Performs the embedding into the Vector DB

Parameters:: j (str) – Parsable JSON string with embedding information

embed_images(data: dict, media: list = None)

Embed dictionary with embedding information.

Parameters:: data (dict) – Dictionary with embedding information