In tpemartin/rmdgrader:

knitr::opts_chunk$set(echo = TRUE, eval=F)
library(drake)
library(econDV)
library(dplyr)

makecondition

library(drake)
library(rmd2drake)
library(dplyr)
library(googleclassroom)
library(readr)
# library(rmdgrader)
library(purrr)
library(rlang)
library(testthat)
library(withr)
rprojroot::is_rstudio_project$make_fix_file() -> .root
.root <- function(){
  "/Users/martinl/Github/course-dashboard-programming-for-data-science"
}
savepath <- 
  file.path(.root(), params$gradingFolder, 
                    params$title)
if(!dir.exists(savepath)) dir.create(savepath)
problemFolder <-
  file.path(.root(), params$submissionFolder, params$title,
            "problem")

Process

# initiate process instance

process <- Process()

# get list of code chunks
process$correctAnsFilename$tryGet_list_codeChunks()
process$inBatch_studentsRmds$tryGet_list_codeChunks()

# If one Rmd content changed, you can update simply that student's respective elements in process instance like:
# process$studentsRmds$HW8_410672033.Rmd$tryGet_list_codeChunks()

Evaluate

After users get list_codeChunks from all Rmd files, it's time to evaluate these codes.

# Initiate an Evaluation Instance
debug(Evaluate)
ei <- Evaluate(process)

# resolve codes in batch
ei$answerValues$batch$resolve$part1()
ei$answerValues$batch$resolve$part2()

# once resolved, answer values will show up for all Rmds such as:
# ei$answerValues$Final_410773111_2.Rmd$values

when screwed

Sometimes students threw in, say totally wrong Rmd; it can break down batch evaluation. In this case, run the following example code (which uses partX) as an example.

for(.it in seq_along(ei$allRmds)){
  cat(glue::glue('{ei$allRmds[[.it]]}\n\n'))
  ei$answerValues[[ei$allRmds[[.it]]]]$resolve$part1()
}

The loop stops at where the problem from, say Final_410973013.Rmd Then you can use the move2problem method to remove it. Just remember to do .nullify after each removal.

process$file.edit$Final_410973013.Rmd()

process$studentsRmds$Final_410973013.Rmd$move2problem()
process$.nullify()

Once done, you can restart a new Evaluation instance (no need to start from Process stage).

Extra bonus: process is actually equipped with file.edit method. Use it like:

process$file.edit$HW7_410672056.Rmd()

individually

# You can also evaluate each Rmd codes one-by-one via their own resolve method. Once resolved, you can see the values after resolution
ei$answerValues$`homework7-ans.Rmd`$resolve$part4()
ei$answerValues$`homework7-ans.Rmd`$resolve$part5()
ei$answerValues$`homework7-ans.Rmd`$resolve$part6()
View(ei$answerValues$`homework7-ans.Rmd`$values)

Save

You only save necessary answer values for grading step via save method.

ei$save(filename = 
          file.path(savepath, "data4step3.Rdata"))

tpemartin/rmdgrader documentation built on Nov. 22, 2022, 6:39 p.m.