Functional Programming Library

Why this library

I was specifically wanting to try using transducers to modify a stream of data generated from an asynchronous iterator.

e.g.

async function(){
    for await (const v of generator) {
        console.log(v)
    }
}

Where the generator is an async generator such as

async function * makeAsyncRangeIterator (start = 0, end = Infinity, step = 1, delay = 10) {
    let iterationCount = 0
    for (let i = start; i < end; i += step) {
        iterationCount++
        yield new Promise((resolve, reject) => {
            setTimeout(() => {
                resolve(i)
            }, delay)
        })
    }
    return iterationCount
  }

or another example would be to process of stream of documents being inserted into a Mongo DB database

async function makeGenerator(config){
  
  const client = await mongoClient.connect(config.MONGO_SERVER);
  const collection = client.db(config.MONGO_DB).collection(config.MONGO_COLLECTION);

  return async function* () {
      const pipeline = [
          { $project : { 
                  "AssetId": "$fullDocument.AssetId", 
                  "Epoch": "$fullDocument.Epoch", 
                  "Value": "$fullDocument.Value", 
                  "SiteId": "$fullDocument.SiteId" } }
      ]
      const changeStream = collection.watch(pipeline,{
          startAtOperationTime: new Date( )
      });
  
      while(true){
          var change = await new Promise( (resolve) =>{
              changeStream.on("change", function(change) {
                  resolve(change)
              });
          })
          delete change._id
          yield change
      }
  }()
}

I found a number of functional programming libraries (tranducer.js, ramda) but found they didnt support asynchronous iterators. They appear to only support synchronous functionality.

How to use

npm install funprog

Then use the functionality that you require (which is visible in index.js)

const { compose, transduceArray, transduceIterator,mapping, filtering } = require('funprog')

Functions

Operators

There are a number of operators which generally take a value and return another value

not,
identity,
isEven,
isGreaterThan,
digitize,
modulus

Stream Transforms

The transforms operate the stream of values

mapping - produce a transformed value for each value in the stream
assign - produce a value with additional properties for each value in the stream
filtering - remove some items from a stream based on a predicate
take - stop generator when a limit is reached
skip - skip values until a threshold is acheived
eventing - combines a set of measurements (single point in time) into events (which have a start/end/duration)
sampling - Sample at a lower frequence than the values are being produced
passthrough - map each event to itself
split - maps from a stream of N events into a stream of N+ events by splitting one event into multiple
randomFilter - randomly includes values at a predefined frequency
neighbors - make n to n + m future stream events available as nth event is processed

Generators

There are a number of functions to produce generators

makeArrayIterator,
makeAsyncRangeIterator,
makeAsyncHasNextRangeIterator

Transducers

There are a set of transducer functions. A transducer is a function taking 4 parameters

a transform,
a reducer,
an initial response ,
a generator

The transducer reads from the generator transforms the value and reduces using the reduce function and initial accumulation

transduceAsyncIterator,
transduceAsyncHasNextIterator,
transduceArray

Composition

Functions that compose functions together

compose - reduce payload processed by array of functions
asyncCompose - reduce payload processed by array of functions returning promises

Tests

Use npm start to run the unit tests

Explanations

Further explanations

bayeslife/funprog