janus.embedding.vectorize

Classes

Vectorizer	Class for creating embeddings/vectors in a specified ChromaDB
VectorizerFactory	Interface for creating a Vectorizer independent of type of ChromaDB client
ChromaDBVectorizer	Factory for Vectorizer that uses ChromaEmbeddingDatabase

Module Contents

class janus.embedding.vectorize.Vectorizer(client, config=None)

Bases: object

Class for creating embeddings/vectors in a specified ChromaDB

Initializes the Vectorizer class

Parameters:

• client (chromadb.Client) – ChromaDB client instance

• config (Optional[Dict[str, Any]]) –

get_or_create_collection(name, model_name=None)

Parameters:

• name (janus.utils.enums.EmbeddingType | str) –

• model_name (Optional[str]) –

Return type:

langchain_community.vectorstores.Chroma

create_collection(embedding_type, model_name=None)

Parameters:

• embedding_type (janus.utils.enums.EmbeddingType) –

• model_name (Optional[str]) –

Return type:

langchain_community.vectorstores.Chroma

collections(name=None)

Parameters:

name (None | janus.utils.enums.EmbeddingType | str) –

Return type:

Sequence[chromadb.Collection]

add_nodes_recursively(code_block, collection_name, file_name)

Embed all nodes in the tree rooted at code_block

Parameters:

• code_block (janus.language.block.CodeBlock) – CodeBlock to embed

• collection_name (janus.utils.enums.EmbeddingType | str) – Collection to add to

• file_name (str) – Name of file containing code_block

Return type:

None

add_text(collection_name, texts, metadatas, ids=None)

Helper function that stores a single text (in an array) and associated metadatas, returning the embedding id

Parameters:

• collection_name (janus.utils.enums.EmbeddingType | str) – Collection to add to

• texts (list[str]) – list of texts to store

• metadatas (list[dict]) – list of metadatas to store

• ids (list[str]) – list of embedding ids (must match lengh of texts), generated if not given by caller

Returns:

list of embedding ids. Raises ValueError if collection not found.

Return type:

list[str]

property config

class janus.embedding.vectorize.VectorizerFactory

Bases: abc.ABC

Interface for creating a Vectorizer independent of type of ChromaDB client

abstract create_vectorizer(path, config={})

Factory method

Parameters:

• path (str | pathlib.Path) –

• config (Dict[str, Any]) –

Return type:

Vectorizer

class janus.embedding.vectorize.ChromaDBVectorizer

Bases: VectorizerFactory

Factory for Vectorizer that uses ChromaEmbeddingDatabase

create_vectorizer(path=Path.home() / '.janus' / 'chroma' / 'chroma-data', config=None)

Parameters:

• path (str | pathlib.Path) – The path to the ChromaDB. Can be either a string of a URL or path or a Path object

• Returns – Vectorizer

• config (Optional[Dict[str, Any]]) –

Return type:

Vectorizer