Flow - Parallel computations on collections

[ShalokShalom]

Computational parallel flows on top of GenStage:

Computational flows with stages.

Flow allows developers to express computations on collections, similar to the Enum and Stream modules, although computations will be executed in parallel using multiple GenStages.

Flow was also designed to work with both bounded (finite) and unbounded (infinite) data. By default, Flow will work with batches of 500 items. This means Flow will only show improvements when working with larger collections. However, for certain cases, such as IO-bound flows, a smaller batch size can be configured through the :min_demand and :max_demand options supported by from_enumerable/2, from_stages/2 and partition/2.

Flow also provides the concepts of “windows” and “triggers”, which allow developers to split the data into arbitrary windows according to event time. Triggers allow computations to be materialized at different intervals, allowing developers to peek at results as they are computed.

Homepage: Flow — Flow v1.2.4
Github: GitHub - dashbitco/flow: Computational parallel flows on top of GenStage

[AstonJ]

Is there anything in particular about Flow that you’d like to discuss Shalok?

I’ve added some further info to your post.

[ShalokShalom]

I simply think that some people are interested.