Docstring mismatch for `Base.download` method about datadeps.jl HOT 4 CLOSED

yes, I think you are correct on all points here.
Having an i_accept flag makes sense as it lets another interface wrap this to get the acceptance some other way.

I'll need to think a little more. But I think that all make sense.
I'm iffy over my use of the function name download at all.
Maybe manually_resolve might be better.

from datadeps.jl.

For the sake of completeness, this is how I currently do it in the PR

function with_accept(f, manual_overwrite)
    auto_accept = if manual_overwrite == nothing
        get(ENV, "DATADEPS_ALWAY_ACCEPT", false)
    else
        manual_overwrite
    end
    withenv(f, "DATADEPS_ALWAY_ACCEPT" => string(auto_accept))
end

function download_dep(depname,
                      dir = DataDeps.determine_save_path(depname);
                      i_accept_the_terms_of_use = nothing)
    datadep = DataDeps.registry[depname]
    with_accept(i_accept_the_terms_of_use) do
        DataDeps.download(datadep, dir)
        nothing
    end
end

The intended purpose is to always trigger the download, even if the dataset is found. Which it currently happily does. I am guessing the rename to manually_resolve you propose doesn't imply changing the behaviour, right?

from datadeps.jl.

I think the name i_accept_the_terms_of_use is not quiet right.
In that the Y/N is not just terms of use, it is also "I accept that this will be downloaded",
and further some things (that are public domain for example) may have no terms of use.

from datadeps.jl.

I guess though it does really get the message across, so it will do for now.

from datadeps.jl.