searchium 0.0.1rc5

Clients of Searchium cloud platform

FVS - is fast vector search

get more info https://www.searchium.ai/

supported methods

• validate_allocation(self) -> bool
• get_loaded_dataset(self) -> List[LoadedDataset]:
• get_datasets -> List[Dataset]
• delete_dataset(dataset_id: str) -> bool
• load_dataset(self, dataset_id: str) -> bool
• unload_dataset(self, dataset_id: str) -> bool
• search(self, dataset_id: str, query: List[List[float]], topk: int = 5) -> SearchResponse
• create_dataset(dataset_id: Optional[str] = None) -> CreateDatasetResponse
• add_chunk(dataset_id: str, list_documents: List[dict]) -> Response
• train_dataset(dataset_id: str) -> Response
• train_status(dataset_id: str) -> TrainStatus
• get_dataset(dataset_id: str) -> Dataset
• delete_document(dataset_id: str, document_id: List[str] = None, delete_all: bool = False) -> ResponseDeleteDocument

example:

import searchium

searchium.init("your_allocation_id", "your_url")

dataset_id = 'your_dataset_uuid'

searchium.create_dataset(dataset_id)

searchium.add_chunk(dataset_id: str, [{"document_id": str=doc_id, "vector": List[float]=embedding, "metadata": dict=metadata}])

searchium.train_dataset(dataset_id)

searchium.load_dataset(dataset_id)

searchium.search(dataset_id: str, query: List[List[float]], topk: int = 5)

searchium.unload_dataset(dataset_id)