These docs are outdated. Read the latest.
PaveDB Python SDK API Reference
This reference is generated from pavesdk.__all__, the SDK public API boundary.
pavesdk.BaseClient
Transport-neutral client surface shared by HTTP and local providers.
BaseClient.close
close(self) -> None
Release resources held by this client.
BaseClient.create_collection
create_collection(
self,
name: str,
*,
tenant: str | None = None,
display_name: str | None = None,
embedder_type: str | None = None,
embed_model: str | None = None,
embedder_config: Mapping[str, Any] | None = None,
) -> Collection
Create a collection and return its tenant-scoped handle.
BaseClient.collection
collection(self, name: str, *, tenant: str | None = None) -> Collection
Return a handle for an existing collection.
BaseClient.list_collections
list_collections(self, *, tenant: str | None = None) -> list[Collection]
List collection handles for a tenant.
BaseClient.list_tenants
list_tenants(self) -> list[str]
List tenants visible to this client.
BaseClient.embedders
embedders(self, *, tenant: str | None = None) -> JsonMap
Return the configured embedders for a tenant.
BaseClient.delete_collection
delete_collection(self, name: str, *, tenant: str | None = None) -> JsonMap
Delete a collection and return the server response.
BaseClient.dump_archive
dump_archive(self, path: str | os.PathLike[str] | None = None) -> Any
Return an archive, or write it to path and return that path.
BaseClient.restore_archive
restore_archive(self, archive_bytes: bytes) -> JsonMap
Restore an archive payload and return the server response.
pavesdk.Collection
Handle for one tenant-scoped PaveDB collection.
Collection(client: BaseClient, tenant: str, name: str)
Create a handle for an existing tenant-scoped collection.
Collection.ingest
ingest(
self,
file: str | os.PathLike[str],
*,
docid: str | None = None,
metadata: Metadata | None = None,
csv_options: Mapping[str, Any] | None = None,
) -> JsonMap
Upload a document file to this collection.
Collection.add
add(
self,
text: str | None = None,
*,
vector: list[float] | None = None,
docid: str | None = None,
metadata: Metadata | None = None,
) -> JsonMap
Add text or a precomputed vector to this collection.
Collection.add_many
add_many(self, documents: list[object]) -> JsonMap
Add a batch of text or vector documents to this collection.
Collection.search
search(
self,
q: str | None = None,
k: int = 5,
*,
vector: list[float] | None = None,
filters: FilterSpec | None = None,
include_common: bool | None = None,
) -> list[JsonMap]
Search this collection by text or a precomputed vector.
Collection.get
get(self, docid: str) -> JsonMap
Return one document by its document ID.
Collection.list_documents
list_documents(self) -> list[JsonMap]
List documents in this collection.
Collection.delete
delete(self, docid: str) -> JsonMap
Delete one document by its document ID.
Collection.detail
detail(self) -> JsonMap
Return this collection’s metadata and summary.
Collection.list_chunks
list_chunks(self, docid: str) -> list[JsonMap]
List indexed chunks for a document.
Collection.get_chunk
get_chunk(self, rid: str) -> JsonMap
Return one indexed chunk by its record ID.
Collection.get_chunk_content
get_chunk_content(self, rid: str) -> JsonMap
Return the raw content and content type for one chunk.
Collection.queries
queries(self, limit: int = 50, offset: int = 0) -> list[JsonMap]
List recorded queries for this collection.
Collection.get_query
get_query(self, qid: str) -> JsonMap
Return one recorded query by its ID.
Collection.replay
replay(self, qid: str) -> list[JsonMap]
Replay a recorded query and return its current matches.
Collection.rename
rename(self, new_name: str) -> Collection
Rename this collection and return the same handle.
Collection.update
update(self, *, display_name: str) -> JsonMap
Update this collection’s display name.
pavesdk.Conflict
Request conflicts with existing state.
pavesdk.HttpClient
Synchronous client for the PaveDB /v1 HTTP API.
HttpClient(
base_url: str,
*,
api_key: str | None = None,
tenant: str = 'default',
timeout: float | httpx.Timeout = 30.0,
headers: Mapping[str, str] | None = None,
http_client: httpx.Client | None = None,
)
Create a client with optional bearer authentication and transport.
HttpClient.close
close(self) -> None
Close the HTTP connection pool when this client owns it.
pavesdk.InvalidRequest
Request shape or arguments are invalid.
pavesdk.LocalClientUnavailable
A local target was requested without a local provider installed.
pavesdk.NotFoundError
Requested resource was not found.
pavesdk.PaveError
Base PaveDB client error.
pavesdk.PAVEDB_API_PREFIX
'/v1'
pavesdk.PAVEDB_API_VERSION
'v1'
pavesdk.Unavailable
The server or an underlying dependency is unavailable.
pavesdk.__version__
Installed pavedb-sdk distribution version.
pavesdk.batch_item
batch_item(document: object) -> JsonMap
Normalize one text or vector item for Collection.add_many.
pavesdk.connect
connect(
target: str | os.PathLike[str] | None = None,
*,
api_key: str | None = None,
tenant: str = 'default',
timeout: float | httpx.Timeout = 30.0,
) -> HttpClient | Any
Connect to an HTTP server or an installed local PaveDB provider.
Omit target for an ephemeral local store. HTTP URLs return HttpClient.