GraphServer

Bases: object

A class for defining and running a Raphtory GraphQL server

Parameters:

Name	Type	Description	Default
work_dir	str | PathLike	the working directory for the server	required
cache_capacity	int	the maximum number of graphs to keep in memory at once	required
cache_tti_seconds	int	the inactive time in seconds after which a graph is evicted from the cache	required
log_level	str	the log level for the server	required
tracing	bool	whether tracing should be enabled	required
tracing_level	str	tracing verbosity (e.g. "ERROR", "WARN", "INFO", "DEBUG", "TRACE").	required
otlp_agent_host	str	OTLP agent host for tracing	required
otlp_agent_port	str	OTLP agent port for tracing	required
otlp_tracing_service_name	str	The OTLP tracing service name	required
config_path	str | PathLike	Path to the config file	required
auth_public_key	str	Base64-encoded public key used to verify bearer tokens	required
require_auth_for_reads	bool	Require auth tokens for read queries	required
create_index	bool	Build a search index on startup	required
heavy_query_limit	int	Maximum number of expensive traversal queries (outComponent, inComponent, edges, outEdges, inEdges, neighbours, outNeighbours, inNeighbours) allowed to run simultaneously. Extra queries are parked on a semaphore.	required
exclusive_writes	bool	If True, ingestion/write operations run one at a time and block reads until complete.	required
disable_batching	bool	If True, batched GraphQL requests are rejected. Prevents bypassing per-request depth/complexity limits.	required
max_batch_size	int	Caps the number of queries accepted in a single batched request.	required
disable_lists	bool	If True, bulk list endpoints on collections are disabled. Clients must use page instead.	required
max_page_size	int	Maximum page size allowed on paged collection queries.	required
max_query_depth	int	Maximum nesting depth of a query.	required
max_query_complexity	int	Maximum estimated cost of a query, based on the number of fields selected.	required
max_recursive_depth	int	Internal safety limit to prevent stack overflows from pathologically structured queries (async-graphql default is 32).	required
max_directives_per_field	int	Maximum number of directives on any single field.	required
disable_introspection	bool	If True, schema introspection is disabled entirely.	required
permissions_store_path	str | PathLike	Path to the permissions store (used by the optional auth extension).	required

__new__(work_dir, cache_capacity=None, log_level=None, tracing=None, tracing_level=None, otlp_agent_host=None, otlp_agent_port=None, otlp_tracing_service_name=None, auth_public_key=None, require_auth_for_reads=None, config_path=None, create_index=None, heavy_query_limit=None, exclusive_writes=None, disable_batching=None, max_batch_size=None, disable_lists=None, max_page_size=None, max_query_depth=None, max_query_complexity=None, max_recursive_depth=None, max_directives_per_field=None, disable_introspection=None, permissions_store_path=None)

Create and return a new object. See help(type) for accurate signature.

run(port=None, timeout_ms=180000)

Run the server until completion.

Parameters:

Name	Type	Description	Default
port	int	The port to use. If not specified, tries 1736 by default and if that is not available starts on an arbitrary port. If specified and the port is in use, the server will fail to start.	None
timeout_ms	int	Timeout for waiting for the server to start. Defaults to 180000.	180000

Returns:

Type	Description
None

start(port=None, timeout_ms=5000)

Start the server and return a handle to it.

Parameters:

Name	Type	Description	Default
port	int	the port to use. If not specified, tries 1736 by default and if that is not available starts on an arbitrary port. If specified and the port is in use, the server will fail to start.	None
timeout_ms	int	wait for server to be online. Defaults to 5000.	5000

The server is stopped if not online within timeout_ms but manages to come online as soon as timeout_ms finishes!

Returns:

Type	Description
RunningGraphServer	The running server

turn_off_index()

Turn off index for all graphs.

Returns:

Type	Description
None

vectorise_all_graphs(embeddings, nodes=True, edges=True)

Vectorise all graphs in the server working directory.

Parameters:

Name	Type	Description	Default
embeddings	OpenAIEmbeddings	the embeddings to use	required
nodes	bool | str	if nodes have to be embedded or not or the custom template to use if a str is provided. Defaults to True.	True
edges	bool | str	if edges have to be embedded or not or the custom template to use if a str is provided. Defaults to True.	True

Returns:

Type	Description
None

vectorise_graph(name, embeddings, nodes=True, edges=True)

Vectorise the graph name in the server working directory.

Parameters:

Name	Type	Description	Default
name	list[str]	the name of the graph to vectorise.	required
embeddings	OpenAIEmbeddings	the embeddings to use	required
nodes	bool | str	if nodes have to be embedded or not or the custom template to use if a str is provided. Defaults to True.	True
edges	bool | str	if edges have to be embedded or not or the custom template to use if a str is provided. Defaults to True.	True

Returns:

Type	Description
None