download_zenodo(): more features

[florisvdh]

New message is now: Will download 11 files (total size: 1.4 GiB) from DOI: .........

[ThierryO]

How strong are the benefits of this new feature compared to the cost of having yet another dependency?

[florisvdh]

The benefit of this information is not neglible for larger datasets.

An alternative could be to copy over the function code of humanize into inborutils, but that seems not so elegant. The humanize package itself is small. Maybe there are alternative packages.

[florisvdh]

@ThierryO , FWIW I've just been looking for alternative functions (present on my own installation :-)). Found no 'common package' dependency for this, though - even units has no function to automatically return human-readable output. Is depending on R.utils package more convenient?

> R.utils::hsize(78945)
[1] "77.1 KiB"
> gdata::humanReadable(78945)
[1] "77.1 KiB"
> humanize::natural_size(78945, suffix_type="binary")
77.1 KiB
> prettyunits::pretty_bytes(78945)
[1] "78.94 kB"
> fs::fs_bytes(78945)
77.1K> 

Only the first three examples are able to present the binary (as opposed to decimal) representation (which I believe is the common standard to be taken) as a character.

[florisvdh]

To make the latter ('character') clear for the less obvious cases:

> class(humanize::natural_size(78945))
[1] "glue"      "character"
> class(fs::fs_bytes(78945))
[1] "fs_bytes" "numeric" 

Also, it can be seen that fs::fs_bytes() does not use the IEC symbol of the kibibyte.

[florisvdh]

@hansvancalster @ThierryO modifications and extras:

• dropped humanize dependency; now uses internal function human_filesize()
• tweaked messages
• added argument quiet to suppress messages (default: FALSE)

[hansvancalster]

Looks OK to me, just one extra assert_that needed.
Only looked at the code, not tested.

[hansvancalster]

OK for me, but will await @ThierryO review before merging.

[ThierryO]

Overall this looks good. Just a few minor comments.

[florisvdh]

Thanks @ThierryO for the extra suggestions. Will look at remaining topics next week!