Subshells: Bringing Multithreading To Jupyter Kernels - Ian Thomas, QuantStack скачать в хорошем качестве

Subshells: Bringing Multithreading To Jupyter Kernels - Ian Thomas, QuantStack

Subshells: Bringing Multithreading To Jupyter Kernels - Ian Thomas, QuantStack Subshells solve the problem of how to interact with a Jupyter kernel whilst it is busy executing code. A subshell is a separate thread of execution which can run code independently of the main shell and shares the same memory space. Using subshells you can inspect the kernel state whilst it is busy executing code, visualise intermediate results before a final result is computed, and execute arbitrary code in parallel. This talk will explain: The problem we are trying to solve. History of the subshells proposal. How subshells work. The reference implementation in ipykernel. Live demonstration of subshells in action using JupyterLab, ipykernel and ipywidgets. Future problems that still need to be solved. For kernel developers, how to implement support for subshells in a language kernel. The talk will be of interest to Jupyter users who will benefit from the use of subshells in their workflows, and kernel developers who wish to implement support for subshells in other language kernels.