The goal of heapsofpapers is to make it easy to respectfully get, well, heaps of papers (and CSVs, and websites, and similar). For instance, you may want to understand the state of open code and open data across a bunch of different pre-print repositories, e.g. Collins and Alexander, 2021, and in that case you need a way to quickly download thousands of PDFs.

Essentially, the main function in the package, heapsofpapers::get_and_save() is a wrapper around a for loop and utils::download.file(), but there are a bunch of small things that make it handy to use instead of rolling your own each time. For instance, the package automatically slows down your requests, lets you know where it is up to, and adjusts for papers that you’ve already downloaded.

## Installation

You can install heapsofpapers from GitHub with:

# install.packages("devtools")
devtools::install_github("RohanAlexander/heapsofpapers")

## Example

Here is an example of getting two papers from SocArXiv, using the main function heapsofpapers::get_and_save():

library(heapsofpapers)
two_pdfs <-
tibble::tibble(
save_here = c("competing_effects_on_the_average_age_of_infant_death.pdf",
)

heapsofpapers::get_and_save(
data = two_pdfs,
save_names = "save_here"
)

By default, the papers are downloaded into a folder called ‘heaps_of’. You could also specify the directory, for instance, if you would prefer a folder called ‘inputs’. Regardless, if the folder doesn’t exist then you’ll be asked whether you want to create it.

heapsofpapers::get_and_save(
data = two_pdfs,
save_names = "save_here",
dir = "inputs"
)

Let’s say that you had already downloaded some PDFs, but weren’t sure and didn’t want to download them again. You could use heapsofpapers::check_for_existence() to check.

heapsofpapers::check_for_existence(data = two_pdfs,
save_names = "save_here")

If you already have some of the files then heapsofpapers::get_and_save() allows you to ignore those files, and not download them again, by specifying that dupe_strategy = "ignore".

heapsofpapers::get_and_save(
data = two_pdfs,
save_names = "save_here",
dupe_strategy = "ignore"
)

• Allow the user to specify how long they would like to wait.
• Add so that it checks whether the file exists and asks the user whether they’d like to re-download that or skip that.
• Add option to save to AWS bucket
• Add something that automatically looks at whether the folder exists and if not, creates it.
• Don’t sleep after the last paper.
• Update check_for_existence()
• Add vignettes of doing this for CSVs and for html.
• Make the printing of the message again optional, or every X or similar, as specified by the user.
• If the directory doesn’t exist, ask if the user wants it to be created and if yes then create it.
• some kind of progress indicator would be great in the console messages (or an option to enable something like that). often, if I’m scraping hundreds of files, I want to know how far into it I am
• for the delay “wiggling” (great call, I love that functionality in wget), may be worth varying it by an increment of the size of the delay. so if i set a second delay, I’d expect the wobble to be of maybe up to 20-25%? if it’s 5sec, that’d be up to a second wobble in either direction.
• Make the behaviour more sophisticated when the file doesn’t exist.
• Update the functionality around dupes.
• Add example around how you can pipe to this e.g. pdfs %>% get_and_save(…)
• Add message that estimates how long it’ll take and asking whether the user would like to proceed.
• Check and develop the s3 behaviour.
• Make the length of the pause dependent on the size of the file that is downloaded, by default.
• Add a check for any PDFs that are very small (which usually indicates there’s something wrong with them).
• Add tests of whether the internet is on.
• Option to save to Dropbox.