pyRTC.Pipeline

Shared-memory transport and hard-RTC process helpers for pyRTC.

This module contains the infrastructure that lets pyRTC components exchange frames and command vectors through named shared-memory blocks, optionally mirror those blocks onto CUDA tensors for compatible deployments, and launch hardware- facing child processes that communicate with the main RTC over a small localhost-based JSON protocol.

Functions

clear_shms(names)
gpu_torch_available()
initExistingShm(shmName[, gpuDevice])
launchComponent(component, confKey[, start])
normalize_gpu_device(gpuDevice[, context])
work(obj, functionName, affinity)	The main working thread for the any Pipeline object

Classes

ImageSHM(name, shape, dtype[, gpuDevice, ...])	Named shared-memory array with metadata and optional GPU mirror state.
Listener(hardware, port)	Server-side control socket for a launched hardware object.
hardwareLauncher(hardwareFile, configFile, port)	Launch and supervise a hardware-side child process.