Search

MessageSearchResult dataclass

content instance-attribute

metadata instance-attribute

qdrant_id instance-attribute

score instance-attribute

__init__(qdrant_id, content, metadata, score)

as_artifact()

as_content()

MessageSearchResults dataclass

results instance-attribute

__init__(results)

__post_init__()

as_artifact()

as_content()

search_messages(uow, query_text, user_id, conversation_uid=None, thread_uid=None, embed_type=None, max_results=None, max_message_age=None, message_types=None, result_index_offset=None, include_long_content=False) async

Search for message contents/snippets in the vectorstore.

Optionally, filter or customize the search.

PARAMETER	DESCRIPTION
uow	Search unit of work. TYPE: MessageSearchUnitOfWork
query_text	Text to search for (by embedding and doing similarity search). TYPE: str
user_id	User id to filter by. TYPE: UserID | None
conversation_uid	Conversation uid to filter by. TYPE: list[UUID] | UUID | None DEFAULT: None
thread_uid	Thread uid to filter by. (Assistant run) TYPE: UUID | None DEFAULT: None
embed_type	Type of embedding to search for (default None means all types). TYPE: MessageEmbeddingType | None DEFAULT: None
max_results	Max number of results to return. TYPE: int | None DEFAULT: None
max_message_age	Maximum age of messages to search. TYPE: timedelta | None DEFAULT: None
message_types	Types of messages to search. TYPE: list[MessageRoles] | None DEFAULT: None
result_index_offset	Offset of the first result to return. TYPE: int | None DEFAULT: None
include_long_content	Whether to also search messages with long content (EmbeddedContentLocation). TYPE: bool DEFAULT: False

RETURNS	DESCRIPTION
MessageSearchResults	MessageSearchResults (has list of MessageSearchResult with methods to transform)