Get error message Remote 'gistr' already exists., then only uploads a blank file.

Check this, then fail, or find some way around

• allow passing in file name, e.g., for geojson, need to have geojson name to render as geojson
• export all extensions

https://developer.github.com/v3/gists/#truncation

Applies for GET requests. if files too large, they are truncated, and only part of the file given back, and a field truncated: TRUE. Here's the bit on truncation:

The Gist API provides up to one megabyte of content for each file in the gist. Every call to retrieve a gist through the API has a key called truncated. If truncated is true, the file is too large and only a portion of the contents were returned in content. If you need the full contents of the file, you can make a GET request to the URL specified by raw_url. Be aware that for files larger than ten megabytes, you’ll need to clone the gist via the URL provided by git_pull_url.

@cpsievert @jennybc @karthik @cboettig if you get a second, could you test out the git functionality (vs. using the github http api)?

devtools::install_github("ropensci/gistr@git2r")
library("gistr")
... # do stuff

see examples https://github.com/ropensci/gistr/blob/git2r/R/gist_create_git.R#L7-L29

in the git2r branch we're using git2r to push to a gist (not a github repo), since gists are git repos just as github repos are. This what mbostocks gistup node.js CLI tool does, using git instead of the http api.

p.s. maybe I should add git functionality to gist_create()

would have to add knitr as an import

What's the easiest path to get this Gist into my hands using gistr?

https://gist.github.com/expersso/4ac33b9c00751fddc7f8#file-google_maps_with_rvest

I tried this

> browse("https://gist.github.com/expersso/4ac33b9c00751fddc7f8#file-google_maps_with_rvest")
Error: Not Found

I'd like to fork it or get that code on my clipboard or get in a local file. It's just an example of wanting to get someone else's Gist and play with it.

Needless to say, I think this is just a documentation gap. And I obviously know how to complete the task other ways, just thought I'd try to do it with gistr for kicks.

from #59

see ropensci/git2r#131

Sorry this isn't really a minimal example, but I'm not sure of a good way to reduce it:

library(git2r)
library(gistr)
repo <- clone("https://github.com/cpsievert/AnimintTest.git", "AnimintTest")
fs <- dir("AnimintTest/FluView", full.names = TRUE)
gist_create(fs, description = "FluView")

#> Error in list2gist(x) : Not coerceable to a gist
#> In addition: Warning message:
#>   In process(response) :
#>   client error: (422) Unprocessable Entity (WebDAV; RFC 4918)

On gist_create() with large files, at least 1.9 MB or larger (the one I found this on), github api returns a 405 - Method not allowed and html error body

But the same file I can paste into browser to create a gist.

And same file uploaded to make a gist using an npm cli tool works fine too - but they use git, not the http api, so not a fair comparison

Weird, maybe it's something going on with httr

Either coerce all objects to json - then when pulled down, jsonlite can convert them back to R objects,

OR

some objects can convert to markdown, like data.frame/matrix to a markdown table

Riffing on this example from gist_create():

file <- system.file("examples", "stuff.Rmd", package = "gistr")
gist_create(file, knit=TRUE, include_source=TRUE)

I define an R script like so:

my_code <-"
#' ---
#' output: md_document
#' ---

#' I shall generate random numbers.
x <- rnorm(1000)

#' And I shall summarize them.
summary(x)"
#EDITED LATER! commenting out my initial garbled code and inserting what I meant :(
#my_code_file <- writeLines(my_code, file.path("~", "tmp", "foo.R"))
my_code_file <- file.path("~", "tmp", "foo.R")
writeLines(my_code, my_code_file)

and I would expect something like this to work:

gist_create(my_code_file, description = "My first R code gist via gistr",
            knit = TRUE, include_source = TRUE)

But instead I get this:

> gist_create(my_code_file, description = "My first R code gist via gistr",
+             knit = TRUE, include_source = TRUE)
Error in readLines(if (is.character(input2)) { : 
  object 'input2' not found

Same result if I remove the YAML front matter, so that's not the problem.

can only do this via git, we could possibly import git2r, then at least temporarily initialize a git repo in the directory in question, upload the binary files after making the gist, then deleting the .gist/ dir (if user wants it deleted).

See http://stackoverflow.com/a/9855131/1091766
gistup may be of help https://developer.github.com/v3/gists/#create-a-gist

from #47

e.g., if a file is knitted, and creates figures, those currently aren't uploaded

Suggest users use imgur, or they could toggle parameter to inject imgur code into .rmd file

@cpsievert FYI, i'm taking the git fxn out for now, putting on another branch, so that I can continue with the cran submission of the remaining changes.

E.g.,

when running gist_create(file, knit=TRUE), toggle parameter to also throw up the source file, either .Rmd or .Rnw

from @jennybc

for example, user could have deleted the gist, in which case gist_create_git() fails

use git2r::remote_ls() ropensci/git2r#172 once next version is on CRAN, but for now not sure what to do

When I do something like this:

gists(what = "minepublic")[[1]] %>%   add_files("some_file.R") %>% edit

my action is mostly updating the gist itself. So shouldn't the verb be update instead of edit?

• Add egs for non-piping workflow

I would like to post a PNG to a gist but I am getting the following error

> gist_create("/tmp/RtmpQZytoH/file2f9877e26e34/thumbnail.png")
Loading required package: rjson
Error in content(response)$html_url (from gist.R#78) : 
  $ operator is invalid for atomic vectors
In addition: Warning messages:
1: client error: (400) Bad Request 
2: In library(package, lib.loc = lib.loc, character.only = TRUE, logical.return = TRUE,  :
  there is no package called ‘rjson’

the image I tried to upload is below -- I uploaded it using git clone, git add, git commit, git push .

https://gist.githubusercontent.com/tdhock/1f01f3359f60582a67cf/raw/19504d4bdf7daa8bdec14868723ef88ded794d79/thumbnail.png

also I can upload lots of other files (text files). Is there some problem with uploading binary files?

> sessionInfo()
R version 3.1.0 (2014-04-10)
Platform: x86_64-unknown-linux-gnu (64-bit)

locale:
 [1] LC_CTYPE=en_CA.UTF-8       LC_NUMERIC=C              
 [3] LC_TIME=en_CA.UTF-8        LC_COLLATE=en_CA.UTF-8    
 [5] LC_MONETARY=en_CA.UTF-8    LC_MESSAGES=en_CA.UTF-8   
 [7] LC_PAPER=en_CA.UTF-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_CA.UTF-8 LC_IDENTIFICATION=C       

attached base packages:
[1] graphics  utils     datasets  grDevices stats     methods   base     

other attached packages:
 [1] RSelenium_1.3.1    XML_3.98-1.1       caTools_1.16       gistr_0.0.2       
 [5] animint_2014.5.22  proto_0.3-10       RColorBrewer_1.0-5 lattice_0.20-29   
 [9] plotly_0.4         ggplot2_0.9.3.1.99 RJSONIO_1.0-3      RCurl_1.95-4.1    
[13] bitops_1.0-6      

loaded via a namespace (and not attached):
 [1] assertthat_0.1   colorspace_1.2-4 compiler_3.1.0   dichromat_2.0-0 
 [5] digest_0.6.4     evaluate_0.5.5   formatR_0.10     grid_3.1.0      
 [9] gtable_0.1.2     httr_0.2         knitr_1.5.34     labeling_0.2    
[13] MASS_7.3-31      munsell_0.4.2    plyr_1.8.1       Rcpp_0.11.1     
[17] reshape2_1.2.2   scales_0.2.3     stringr_0.6.2    tools_3.1.0     

Right now, running code is only in gist_create(), but not avail if you want to edit an existing gist.

Allow knitting on an update in edit():

gists('minepublic')[[1]] %>% edit(knit  = TRUE)

or perhaps as a separate step

gists('minepublic')[[1]] %>% knit() %>% edit()

from #47 e.g.,

- 4ac33b9c00751fddc7f8 -- # folder
                       | - file_name.R # file
                       | - file_name.md # file

@hadley @ramnathv

I've reworked the API for the package on the dev branch.

I'd love to know what you think.

Hadley, I've more or less copied the interface you introduced in analogsea as I thought it would work well in the gist context.

See the readme for intro: https://github.com/ropensci/gistr/tree/dev#quick-start

borking because json keys have changed slightly, make this more robust to chagnes on their end, push chagne to cran

obj <- mtcars[1:5,]
gist_create(obj)

obviously more complicated sets of objects should be possible

It'd be nice to use devtools::install_github for installation which requires a DESCRIPTION.

should help if users have yaml frontmatter

from #29

The gist class is extremely frustrating to work with. Unless I'm doing something wrong, the actual gist ids are inaccessible. e..g.

> x<- gists("minepublic")
> x[[1]]
<gist>44eb9788a675d42e6dd5
  URL: https://gist.github.com/44eb9788a675d42e6dd5
  Description: 
  Public: TRUE
  Created/Edited: 2015-02-16T09:11:54Z / 2015-02-16T09:11:54Z
  Files: temp.geojson

I would love to just have a simple data.frame (or list that can be easily flattened) to work with for a practical use case other than uploading. like

gist("minepublic") %>%
 filter(title = "temp") %>%
delete 

e.g., this one https://gist.github.com/sckott/1548267c71d267706de0 the output file name is not what it should be by default code.R, or rather code.md

not all files in a directory

Can do 2 things with tabl()

• Separate row for each file, or
• concatenate fields where there are multiple unique values into one comma sep string

We basically do the first now, but then other fields aren't filled in for each row, and when combined with other gists, means the gist id rows belong with are lost.

I think option 2 might be cleaner, but not sure

@karthik sorry to keep asking questions, but it looks like the file name for this fxn is only needed to append to the url (that is, no local file manipulation needed in the fxn). Can we just get that from the output of gist_create()?, and if so, then the output of gist_create() can be piped to gist_map() as it's first argument? Or are you trying to support if someone wants to use a local file too?

e.g., #41

In some cases good error messages are returned from the github api, pass these on

Error status code come before header check, and maybe stop instead of warn

I am experimenting with an implementation of gistr using reference classes and was wondering if this API would be cleaner. You will find my preliminary implementation here.

The idea is to allow a workflow that resembles this

# set options github.username and github.password
# alternately pass username and password as arguments
gist = Gist()

# get
gist$get('public') # get public gists
gist$get('starred') # get starred gists
gist$get(user = 'hadley') # get gists of a user
gist$get("d5029341cfa9c2d02fa9") # get a gist by id

# add_files and post
gist$add_files(path = ".")
gist$post("My First Gist")

# make some edits to the files you added to the gist
# fill out the gist_id (can be automated)
gist_id = 
gist$update(id = gist_id)

# delete files
# specific files
gist$delete(gist_id, files = files_to_delete)

# delete complete gist
gist$delete

One big advantage I see to using reference classes here is that we can implement pretty sophisticated stuff taking advantage of the fact that ref classes keep track of state. Another advantage is that the API would mirror the official gist API very closely, making it easier to adapt to any changes.

I look forward to comments/suggestions on using this approach.

blockbuilder.org is a very nice online gist editor primarily built for d3.js bl.ocks, but it could easily be used as an online gist editor for anything. Would it be worthwhile to add a function to open a gist in blockbuilder?

I am running into an issue with creating gists with unicode characters.

The characters are mangled when I try this code:

library(curl)
d3 <- "https://github.com/mbostock/d3/archive/v3.0.6.zip"
tmp <- tempfile()
curl::curl_download(d3, tmp, mode = "wb")
unzip(tmp)
unlink(tmp)

library(gistr)
gistr::gist_create(files = "d3-3.0.6/d3.js")

This might be a Windows issue because the same code works fine for @cpsievert .

> sessionInfo()
R version 3.1.3 (2015-03-09)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 8 x64 (build 9200)

locale:
[1] LC_COLLATE=English_United States.1252 
[2] LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252
[4] LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] gistr_0.1.0 curl_0.5   

loaded via a namespace (and not attached):
 [1] bitops_1.0-6    digest_0.6.8    evaluate_0.5.5  formatR_1.0    
 [5] httpuv_1.3.2    httr_0.6.1      jsonlite_0.9.14 knitr_1.9      
 [9] magrittr_1.5    R6_2.0.1        Rcpp_0.11.4     RCurl_1.95-4.5 
 [13] stringr_0.6.2   tools_3.1.3 

✔️ into it

ropensci/git2r#116