split off download functionality for use_course()?

[beanumber]

I see the value of use_course(), but I sometimes want to download a ZIP file that is not an RStudio project, and use it. For example, a ZIP file that lives somewhere online that contains some shapefiles that I want to use.

So how would you feel about a PR that includes use_zip(), which is just like use_course() except that it returns an fs_path to the unzipped directory? use_course() could would then call use_zip() to maintain its current functionality.

[jennybc]

I see the value of use_course(), but I sometimes want to download a ZIP file that is not an RStudio project

This already works. Just use it on a ZIP archive that does not contain an Rproj file.

For example, try this: use_course("rstd.io/wtf-explore-libraries")

But I suspect we are misunderstanding each other.

[beanumber]

OK. But indulge me by considering three things:

1. As per use_course() does not accept a ZIP with mime type "application/octet-stream" #304 , use_course() has a narrow scope, so for example this doesn't work:

usethis::use_course("https://www.fueleconomy.gov/feg/epadata/16data.zip")
#> ✔ Downloading 'https://www.fueleconomy.gov/feg/epadata/16data.zip'
#> 
Downloaded: 0.02 MB  (2%)
Downloaded: 0.02 MB  (2%)
Downloaded: 0.03 MB  (3%)
...
Downloaded: 0.93 MB  (100%)
#> Error: Download does not have MIME type 'application/zip'.
#> Instead it's 'application/x-zip-compressed'.

^{Created on 2019-05-31 by the reprex package (v0.3.0)}

2. I find the forced interactivity to be annoying most of the time. I understand why it's there, but can it be optional? Can it take an interactive argument to turn this off?

3. Why is the function called use_course() if I'm using it for something that has nothing to do with courses?

All I'm proposing would be to expose use_zip() as a more general-purpose function (with a fix for (1)), and then use_course() could continue to work as it currently does.

[jennybc]

Is this true? What you like about use_course() and want to extract into use_zip() is a nice download function that:

• Figures out a reasonable file name (vs. making you do that)
• Unpacks the ZIP archive in a sane way (vs. the typical result that often results in 1 more or 1 less level of nesting than you wanted)

[beanumber]

Yes. And returns the fs_path to the resulting directory.

So instead of:

src <- "https://www.fueleconomy.gov/feg/epadata/16data.zip"

download.file(src, destfile = "fueleconomy.zip")
unzip("fueleconomy.zip", exdir = "fueleconomy/")
lcl <- fs::path("fueleconomy/")

We'd just have:

src <- "https://www.fueleconomy.gov/feg/epadata/16data.zip"
lcl <- usethis::use_zip(src)

[jennybc]

OK see what you think of use_zip(). I have also been using use_course() just to get sane ZIP download + unpack behaviour, so I see your point.

[beanumber]

Amazing!

The docs say: "doesn't insist on interactive confirmation.", but tidy_unzip() still wants to confirm the deletion of the ZIP file.

[beanumber]

And thank you!

[jennybc]

still wants to confirm the deletion of the ZIP file.

Yeah I was referring to interaction around destdir. What are your thoughts around deleting the ZIP file? I can't very well delete by default, but deleting is so nice. I think I'm OK with being asked this in an interactive setting.

[beanumber]

Could there be a delete_zip argument that defaults to "ask" but will be quiet if it's set to TRUE or FALSE?

[jennybc]

OK now you can do use_zip("r-lib/rematch2", cleanup = TRUE).

[beanumber]

Thanks @jennybc !