Discussed in networkx/networkx#5446

steps to reproduce

Write a loop, from 1 to 80000, each time add a random int to the max heap.

In theory it takes very little time(NlogN, N=80000, <1sec ), but the program does take a long time.

I'v also tested the BinaryHeap in https://github.com/SolutionsDesign/Algorithmia, it performs well, so it is probably due to the bad algorithm.

https://en.wikipedia.org/wiki/INI_file

Intro

I am getting TypeError: can not serialize 'BaseTextRank' object when trying to use spaCy's multiprocessing in nlp.pipe with a textrank pipeline component.

Sorry if this a known/expected feature/limitation - I couldn't find anything by searching repo. I generally find (spaCy's) multiprocessing a bit temperamental anyhow, but this seems to just not work.

_PS. thanks for all

Feature description (Mandatory)

Please add the capability to specify the type of array instead of defaulting to string. i.e.
apoc.array.int()

How this feature can improve the project?

By default, it is created as a string. Some customers have asked for the ability to hint at the data type.

What is the feature or improvement you would like to see?

A function that computes the graph center, and the graph periphery. The graph center is the set of vertices that has minimal eccentricity. The periphery is the set with maximal eccentricity.

Use cases for the feature

This is a standard concept in graph theory.

**Implementatio