- Embedded
- Python SDK
- JS/TS
CyborgVectorStore(
index_name: str,
index_key: bytes | list[int] | tuple[int, ...] | np.ndarray,
api_key: str,
embedding: Embeddings,
index_location: DBConfig,
config_location: DBConfig,
items_location: Optional[DBConfig] = None,
index_type: str = "ivfsq",
index_config_params: Optional[Dict[str, Any]] = None,
dimension: Optional[int] = None,
metric: str = "cosine",
gpu_config: Optional[GPUConfig] = None
)
Parameters
| Parameter | Type | Description |
|---|---|---|
index_name | str | Name of the index (must be unique) |
index_key | bytes | list[int] | tuple[int, ...] | np.ndarray | 32-byte encryption key for securing index data |
api_key | str | API key for CyborgDB authentication |
embedding | Embeddings | LangChain Embeddings instance |
index_location | DBConfig | Configuration for index data storage location |
config_location | DBConfig | Configuration for index config storage location |
items_location | Optional[DBConfig] | (Optional) Location for item data storage (default: memory) |
index_type | str | Type of index: "ivfflat", "ivfpq", or "ivfsq" (default: "ivfsq") |
index_config_params | Optional[Dict[str, Any]] | (Optional) Additional index configuration parameters |
dimension | Optional[int] | (Optional) Embedding dimension (auto-inferred if not provided) |
metric | str | Distance metric: "cosine", "euclidean", or "squared_euclidean" (default: "cosine") |
gpu_config | Optional[GPUConfig] | (Optional) GPU configuration for accelerated operations |
Returns
CyborgVectorStore: Initialized vector store instanceExample Usage
from cyborgdb_core.integrations.langchain import CyborgVectorStore
from cyborgdb_core import DBConfig
from langchain_openai import OpenAIEmbeddings
# Generate a secure key
index_key = CyborgVectorStore.generate_key(save=True)
# Initialize with LangChain Embeddings
store = CyborgVectorStore(
index_name="my_documents",
index_key=index_key,
api_key="your-api-key",
embedding=OpenAIEmbeddings(),
index_location=DBConfig("rocksdb"),
config_location=DBConfig("rocksdb"),
index_type="ivfsq",
metric="cosine"
)
CyborgVectorStore(
index_name: str,
index_key: bytes,
api_key: str,
embedding: Union[str, Embeddings, SentenceTransformer],
base_url: str,
verify_ssl: Optional[bool] = None,
index_type: str = "ivfflat",
index_config_params: Optional[Dict[str, Any]] = None,
dimension: Optional[int] = None,
metric: str = "cosine"
)
Parameters
| Parameter | Type | Description |
|---|---|---|
index_name | str | Name of the index (must be unique) |
index_key | bytes | 32-byte encryption key for securing index data |
api_key | str | API key for CyborgDB authentication |
embedding | Union[str, Embeddings, SentenceTransformer] | Embedding model name or instance |
base_url | str | Base URL of the CyborgDB microservice endpoint |
verify_ssl | Optional[bool] | (Optional) SSL verification. When None, automatically disabled for localhost and http:// URLs |
index_type | str | Type of index: "ivfflat", "ivfpq", or "ivfsq" (default: "ivfflat") |
index_config_params | Optional[Dict[str, Any]] | (Optional) Additional index configuration parameters |
dimension | Optional[int] | (Optional) Embedding dimension (auto-inferred if not provided) |
metric | str | Distance metric: “cosine”, “euclidean”, or “squared_euclidean” (default: “cosine”) |
Returns
CyborgVectorStore: Initialized vector store instanceThe Python SDK requires a running CyborgDB service. Use
base_url to point to your service endpoint instead of providing DBConfig storage locations.Example Usage
from cyborgdb.integrations.langchain import CyborgVectorStore
# Generate a secure key
index_key = CyborgVectorStore.generate_key(save=True)
# Initialize with string embedding model name
store = CyborgVectorStore(
index_name="my_documents",
index_key=index_key,
api_key="your-api-key",
embedding="sentence-transformers/all-MiniLM-L6-v2",
base_url="http://localhost:8000",
index_type="ivfflat",
metric="cosine"
)
new CyborgVectorStore(embeddings: EmbeddingsInterface, config: CyborgVectorStoreConfig)
Parameters
| Parameter | Type | Description |
|---|---|---|
embeddings | EmbeddingsInterface | LangChain embeddings instance |
config | CyborgVectorStoreConfig | Configuration object (see below) |
CyborgVectorStoreConfig
| Parameter | Type | Description |
|---|---|---|
baseUrl | string | Base URL of the CyborgDB microservice endpoint |
apiKey | string | API key for CyborgDB authentication |
indexName | string | Name of the index (must be unique) |
indexKey | Uint8Array | 32-byte encryption key for securing index data |
indexType | string | (Optional) Type of index: "ivfflat", "ivfpq", or "ivfsq" (default: "ivfflat") |
indexConfigParams | Record<string, number> | (Optional) Additional index configuration parameters |
dimension | number | (Optional) Embedding dimension (auto-inferred if not provided) |
metric | string | (Optional) Distance metric: "cosine", "euclidean", or "squared_euclidean" (default: "cosine") |
verifySsl | boolean | (Optional) SSL verification (default: true) |
Returns
CyborgVectorStore: Initialized vector store instanceThe JS/TS SDK requires a running CyborgDB service. Use
baseUrl to point to your service endpoint instead of providing DBConfig storage locations.Example Usage
import { CyborgVectorStore } from 'cyborgdb';
import { OpenAIEmbeddings } from '@langchain/openai';
const store = new CyborgVectorStore(
new OpenAIEmbeddings(),
{
baseUrl: "http://localhost:8000",
apiKey: "your-api-key",
indexName: "my_documents",
indexKey: CyborgVectorStore.generateKey(),
indexType: "ivfflat",
metric: "cosine"
}
);
Exceptions
ValueError
ValueError
- Throws if index_type is invalid
RuntimeError
RuntimeError
- Throws if no embedding model provided and dimension not specified