zkamvar

#' Produce a list of all GitHub contributors to a project
#'
#' If you want a way to give thanks to specific people who have contributed to a project
#' since a given release, this will extract the GitHub user names from a given package's 
#' NEWS file. This assumes that 1) the package has a `NEWS.md` file and 2) the package 
#' uses level 1 headers and semantic versioning for each version in the news file. 
#' 
#' @param package the name of a package
#' @param since a version number (must match a version recorded in the NEWS)
#' @return a character vector with GitHub user names

baptiste

library(ggforce)
library(purrr)
library(tibble)
library(gganimate)

web_strand <- function(side = 1, bend = 0.5, angle = 0, start = c(0,0)){
  pos <- cbind(x=c(0, 1, 2), y = bend*c(0, 2, 0))
  post <- pos %*% matrix(c(cos(angle), -sin(angle), sin(angle), cos(angle)), ncol=2, byrow=TRUE)
  xt <- post[,1] + start[1]
  yt <- post[,2] + start[2]

jrosell

# GAM model with a local smoother 
library(tidyverse)
set.seed(2)
elapsed <- arima.sim(model = list(order = c(0, 1, 0)), n=200) + 20
elapsed <- pmax(elapsed, 1)
data <- tibble(
    x = 1:201,
    elapsed = elapsed
)
plot(data)

eliocamp

# This is now available into ggpercentogram. 
# https://github.com/eliocamp/ggpercentogram/
StatQuantileBin <- ggplot2::ggproto("StatQuantileBin", ggplot2::StatBin, 
   default_aes = ggplot2::aes(x = ggplot2::after_stat(density), y = ggplot2::after_stat(density), weight = 1),                                   
   compute_group = function(data, scales, 
                            binwidth = NULL, bins = 30, breaks = NULL, trim = 0,
                            closed = c("right", "left"), pad = FALSE,
                            flipped_aes = FALSE,
                            # The following arguments are not used, but must
                            # be listed so parameters are computed correctly

chainsawriot

require(rang)
require(purrr)
require(igraph)

convert_edgelist <- function(x) {
    output <- data.frame(x = x$pkg, y = rang:::.extract_queryable_dependencies(x$original, x$no_enhances, x$no_suggests))
    for (dep in x$deps) {
        if (!rang:::.is_terminal_node(dep, x$no_enhances)) {
            el <- data.frame(x = unique(dep$x_pkgref), y = rang:::.extract_queryable_dependencies(dep, x$no_enhances, x$no_suggests))
            output <- rbind(output, el)

MichaelChirico

library(withr)
library(data.table)
r_dir <- "~/github/r-svn"

all_commits <- with_dir(r_dir, system('git log --pretty="format:%H %b"', intern=TRUE))
all_commits <- all_commits[nzchar(all_commits)]
hashes <- substr(grep("^[0-9a-f]{40} ", all_commits, value=TRUE), 1, 40)

trunk_hash <- head(hashes, 1L)
# grep("trunk@33000", all_commits, value = TRUE)

MichaelChirico

#!/usr/local/bin/Rscript

library(potools)
suppressPackageStartupMessages(library(data.table))

script_wd = setwd("~/github/r-svn")
GIT_COMMIT = if (interactive()) readline('git commit: ') else commandArgs(TRUE)
setwd('src/library')

get_po_messages <- potools:::get_po_messages

moodymudskipper

set_invisible_attr <- function(x, ...) {
  x_chr <- as.character(substitute(x))
  pf <- parent.frame()
  if(bindingIsActive(x_chr, pf)) {
    env <- environment(activeBindingFunction(x_chr, pf))
    args <- list(...)
    env$closure$attrs[names(args)] <- args
    return(x)
  }

TG9541

Using R on a Raspberry Pi Zero W

Using R on an embedded device might sound like a wacky thing to do but for "small scale data analytics" and for learning to analyze real-world data (e.g. using JAGS or STAN) it's a good match.

Yes, compared to a PC it's relatively slow - but consider this:

• it runs on small power budget (maybe 0.5W when idling)
• it has a hardware watchdog which makes it more likely that it can run for many months unattended
• Raspbian-lite is based on Debian and it comes with all kinds of well tested tools and applications

walkerke

library(tidycensus)
library(ggiraph)
library(tidyverse)
library(patchwork)

vt_income <- get_acs(
  geography = "county",
  variables = "B19013_001",
  state = "VT",
  year = 2019,