Learn R Programming

gistr (version 0.9.0)

gist_create_git: Create a gist via git instead of the GitHub Gists HTTP API

Description

Create a gist via git instead of the GitHub Gists HTTP API

Usage

gist_create_git(
  files = NULL,
  description = "",
  public = TRUE,
  browse = TRUE,
  knit = FALSE,
  code = NULL,
  filename = "code.R",
  knitopts = list(),
  renderopts = list(),
  include_source = FALSE,
  artifacts = FALSE,
  imgur_inject = FALSE,
  git_method = "ssh",
  sleep = 1,
  ...
)

Arguments

files

Files to upload. this or code param must be passed

description

(character) Brief description of gist (optional)

public

(logical) Whether gist is public (default: TRUE)

browse

(logical) To open newly create gist in default browser (default: TRUE)

knit

(logical) Knit code before posting as a gist? If the file has a .Rmd or .Rnw extension, we run the file with knit, and if it has a .R extension, then we use render

code

Pass in any set of code. This can be a single R object, or many lines of code wrapped in quotes, then curly brackets (see examples below). this or files param must be passed

filename

Name of the file to create, only used if code parameter is used. Default to code.R

knitopts, renderopts

(list) List of variables passed on to knit, or render

include_source

(logical) Only applies if knit=TRUE. Include source file in the gist in addition to the knitted output.

artifacts

(logical/character) Include artifacts or not. If TRUE, includes all artifacts. Or you can pass in a file extension to only upload artifacts of certain file exensions. Default: FALSE

imgur_inject

(logical) Inject imgur_upload into your .Rmd file to upload files to https://imgur.com/. This will be ignored if the file is a sweave/latex file because the rendered pdf can't be uploaded anyway. Default: FALSE

git_method

(character) One of ssh (default) or https. If a remote already exists, we use that remote, and this parameter is ignored.

sleep

(integer) Seconds to sleep after creating gist, but before collecting metadata on the gist. If uploading a lot of stuff, you may want to set this to a higher value, otherwise, you may not get accurate metadata for your gist. You can of course always refresh afterwards by calling gist with your gist id.

...

Further args passed on to verb-POST

Details

Note that when browse=TRUE there is a slight delay in when we open up the gist in your default browser and when the data will display in the gist. We could have this function sleep a while and guess when it will be ready, but instead we open your gist right after we're done sending the data to GitHub. Make sure to refresh the page if you don't see your content right away.

Likewise, the object that is returned from this function call may not have the updated and correct file information. You can retrieve that easily by calling gist() with the gist id.

This function uses git instead of the HTTP API, and thus requires the R package git2r. If you don't have git2r installed, and try to use this function, it will stop and tell you to install git2r.

This function using git is better suited than gist_create() for use cases involving:

  • Big files - The GitHub API allows only files of up to 1 MB in size. Using git we can get around that limit.

  • Binary files - Often artifacts created are binary files like .png. The GitHub API doesn't allow transport of binary files, but we can do that with git.

Another difference between this function and gist_create() is that this function can collect all artifacts coming out of a knit process.

If a gist is somehow deleted, or the remote changes, when you try to push to the same gist again, everything should be fine. We now use tryCatch on the push attempt, and if it fails, we'll add a new remote (which means a new gist), and push again.

See Also

gist_create(), gist_create_obj()

Examples

Run this code
# NOT RUN {
# prepare a directory and a file
unlink("~/gitgist", recursive = TRUE)
dir.create("~/gitgist")
file <- system.file("examples", "stuff.md", package = "gistr")
writeLines(readLines(file), con = "~/gitgist/stuff.md")

# create a gist
gist_create_git(files = "~/gitgist/stuff.md")

## more than one file can be passed in
unlink("~/gitgist2", recursive = TRUE)
dir.create("~/gitgist2")
file.copy(file, "~/gitgist2/")
cat("hello world", file = "~/gitgist2/hello_world.md")
list.files("~/gitgist2")
gist_create_git(c("~/gitgist2/stuff.md", "~/gitgist2/hello_world.md"))

# Include all files in a directory
unlink("~/gitgist3", recursive = TRUE)
dir.create("~/gitgist3")
cat("foo bar", file="~/gitgist3/foobar.txt")
cat("hello", file="~/gitgist3/hello.txt")
list.files("~/gitgist3")
gist_create_git("~/gitgist3")

# binary files
png <- system.file("examples", "file.png", package = "gistr")
unlink("~/gitgist4", recursive = TRUE)
dir.create("~/gitgist4")
file.copy(png, "~/gitgist4/")
list.files("~/gitgist4")
gist_create_git(files = "~/gitgist4/file.png")

# knit files first, then push up
# note: by default we don't upload images, but you can do that, 
# see next example
rmd <- system.file("examples", "plots.Rmd", package = "gistr")
unlink("~/gitgist5", recursive = TRUE)
dir.create("~/gitgist5")
file.copy(rmd, "~/gitgist5/")
list.files("~/gitgist5")
gist_create_git("~/gitgist5/plots.Rmd", knit = TRUE)

# collect all/any artifacts from knitting process
arts <- system.file("examples", "artifacts_eg1.Rmd", package = "gistr")
unlink("~/gitgist6", recursive = TRUE)
dir.create("~/gitgist6")
file.copy(arts, "~/gitgist6/")
list.files("~/gitgist6")
gist_create_git("~/gitgist6/artifacts_eg1.Rmd", knit = TRUE, 
   artifacts = TRUE)

# from a code block
gist_create_git(code={'
x <- letters
numbers <- runif(8)
numbers

[1] 0.3229318 0.5933054 0.7778408 0.3898947 0.1309717 0.7501378 0.3206379 0.3379005
'}, filename="my_cool_code.R")

# Use https instead of ssh
png <- system.file("examples", "file.png", package = "gistr")
unlink("~/gitgist7", recursive = TRUE)
dir.create("~/gitgist7")
file.copy(png, "~/gitgist7/")
list.files("~/gitgist7")
gist_create_git(files = "~/gitgist7/file.png", git_method = "https")
# }

Run the code above in your browser using DataLab