furly

an R package for blazing fast parsing of JSON urls

The goal of furly is to combine the fast async curl downloads and the blazing fast JSON parsing of RcppSimdJso

Installation

You can install the development version of furly from Github:

devtools::install_github("melsiddieg/furly")

install.packages("furly")

Example

This is a basic example which shows you how to solve a common problem:

library(furly)
url <- "https://bit.ly/cellbase1"
url2 <- "https://bit.ly/cellbase2"
url3 <- "https://bit.ly/cellbase3"
url4 <- "https://bit.ly/cellbase4"
urls <- c(url,url2,url3,url4)
res<-furly(urls)

Benchmarks

bench <- microbenchmark::microbenchmark(
  jsonlite= lapply(urls, function(x)jsonlite::fromJSON(x)),
  RcppSimdJson = RcppSimdJson::fload(urls),
  furly = furly(urls),times = 3
)

Unit: relative
        expr       min        lq      mean    median       uq       max neval
    jsonlite  5.543655  5.839211  4.736492  6.126533  4.45725  3.535009     3
 RcppSimdJson 15.977387 18.850337 15.918384 21.643254 15.89797 12.723832     3
       furly  1.000000  1.000000  1.000000  1.000000  1.00000  1.000000     3