orangeqs.juice.schemas.ipython#

IPython related schemas.

Module Contents#

Classes#

IPythonKernelConfigs

Configurations for IPython kernels of all services.

KernelConnectionInfo

IPython kernel connection of a single service.

KernelConnectionRegistry

Registry of IPython kernel connection information for all services.

API#

class orangeqs.juice.schemas.ipython.IPythonKernelConfigs(/, **data: Any)#

Bases: orangeqs.juice.settings.Configurable

Configurations for IPython kernels of all services.

IPython kernels in OrangeQS Juice services use the default configuration of IPython if not configured here. This means that the kernel will use a random key and ports by default.

This configuration allows to override the default IPython configuration.

filename: ClassVar[str]#

‘ipykernel’

services: dict[str, KernelConnectionInfo]#

‘Field(…)’

IPython kernel configuration for all services.

for_service(service_name: str) → orangeqs.juice.schemas.ipython.KernelConnectionInfo#

Return the connection info configured for a service.

Defaults unset ip/supervisor_port to the address of a service spawned by Juice Orchestrator in the same Podman network (juice-<service_name>, port 9001 - matching ServiceSettings.supervisor_port’s own default).

Parameters#

  • service_name (str): The name of the service.

Returns#

  • (KernelConnectionInfo): The connection info for the service, with ip/supervisor_port defaulted if unset.

class orangeqs.juice.schemas.ipython.KernelConnectionInfo(/, **data: Any)#

Bases: pydantic.BaseModel

IPython kernel connection of a single service.

This schema is compatible with the connection info used by IPython, jupyter_client.connect.KernelConnectionInfo through the property jupyter_client_connection_info.

Also carries supervisor_port, which is not part of the actual IPython kernel connection (excluded from jupyter_client_connection_info) but piggybacked here so IPythonKernelConfigs can serve as a single per-service address config for both a service’s IPython kernel and its supervisord XML-RPC server (see IPythonKernelConfigs.for_service). This makes this schema more a “SupervisedIPythonServiceConnectionInfo” than pure IPython kernel connection info.

shell_port: orangeqs.juice.schemas.common.PortNumber | None#

None

IPython shell port.

iopub_port: orangeqs.juice.schemas.common.PortNumber | None#

None

IPython IOPub port.

stdin_port: orangeqs.juice.schemas.common.PortNumber | None#

None

IPython standard input port.

control_port: orangeqs.juice.schemas.common.PortNumber | None#

None

IPython control port.

hb_port: orangeqs.juice.schemas.common.PortNumber | None#

None

IPython heartbeat port.

key: str | None#

None

Communication secret key.

ip: str | None#

None

IP address service listens, or the hostname that will resolve to IP address.

transport: Literal[tcp, ipc] | None#

None

IPython transport.

signature_scheme: str | None#

None

IPython communication signature scheme.

supervisor_port: orangeqs.juice.schemas.common.PortNumber | None#

None

Port this service’s supervisord inet HTTP server listens on.

Not part of the IPython kernel connection - see the class docstring.

property jupyter_client_connection_info: jupyter_client.connect.KernelConnectionInfo#

Convert to a Jupyter Client KernelConnectionInfo object.

class orangeqs.juice.schemas.ipython.KernelConnectionRegistry(/, **data: Any)#

Bases: orangeqs.juice.settings.RuntimeData

Registry of IPython kernel connection information for all services.

IPython services store their connection information in this registry on start. IPython clients use this registry to load the connection information for the service they want to connect to. The registry is stored in the shared runtime data directory

filename: ClassVar[str]#

‘ipykernel-connection’

services: dict[str, orangeqs.juice.schemas.ipython.KernelConnectionInfo]#

‘Field(…)’

Kernel connection information for all services.