mnemosine.storage.Storage.embed

Kind: method

Signature

Storage.embed(self, node, text=None)

Documentation

Compute and store an embedding for a node.

See mnemosine.embed.embed. Requires embed_fn.

Args:

• node (Node): The node to embed.
• text (str | None): Optional text; defaults to node.content.

Returns:

• list of float: The computed embedding.

Raises:

• EmbeddingRequired: If embed_fn is unset or there is no text.
• VectorError: If the embedding function returns an invalid vector.

Example:

>>> db.embed_fn = lambda t: [1.0, 0.0]
>>> db.embed(doc)
[1.0, 0.0]