typo in documentation?

[Alexander-Barth]

I try to run the example from nice documentation at https://juliaclimate.github.io/ClimateSatellite.jl/dev/tutorials/downloads.html:

julia> using ClimateSatellite, Dates
julia> clisatdownload("mptw2m",Date(2018,1))
[ Info: 2020-07-14T08:22:33.714 - No directory path was given.  Setting to default path: /home/abarth/research/CliSat/mptw2m for ClimateSatellite data downloads.
┌ Warning: 2020-07-14T08:22:33.729 - The path /home/abarth/research/CliSat/mptw2m does not exist.  A new directory will be created here.  Therefore if you already have an existing repository for ClimateSatellite data, make sure that /home/abarth/research/CliSat/mptw2m is the correct location.
└ @ ClimateSatellite ~/.julia/dev/ClimateSatellite/src/frontend.jl:21
[ Info: 2020-07-14T08:22:33.732 - Creating path /home/abarth/research/CliSat/mptw2m ...
ERROR: BoundsError: attempt to access 0-element Array{Any,1} at index [1]
Stacktrace:
 [1] getindex(::Array{Any,1}, ::Int64) at ./array.jl:788
 [2] clisatinfo!(::Dict{Any,Any}, ::String) at /home/abarth/.julia/dev/ClimateSatellite/src/frontend.jl:82
 [3] clisatdownload(::String, ::Date; email::String, path::String, regions::Array{String,1}, overwrite::Bool) at /home/abarth/.julia/dev/ClimateSatellite/src/raw.jl:27
 [4] clisatdownload(::String, ::Date) at /home/abarth/.julia/dev/ClimateSatellite/src/raw.jl:22
 [5] top-level scope at REPL[5]:1
 [6] eval(::Module, ::Any) at ./boot.jl:331
 [7] eval_user_input(::Any, ::REPL.REPLBackend) at /buildworker/worker/package_linux64/build/usr/share/julia/stdlib/v1.4/REPL/src/REPL.jl:86
 [8] run_backend(::REPL.REPLBackend) at /home/abarth/.julia/packages/Revise/AMRie/src/Revise.jl:1023
 [9] top-level scope at none:0

Could it be that the id mptw2m should be spelled mtpw2m in the documentation ?

I am wondering if you considered julia types and dispatch to simplify code like here
https://github.com/JuliaClimate/ClimateSatellite.jl/blob/master/src/raw.jl#L29

Every dataset could be a type which implements a specific download function (common download function could be handled by abstract types).

dataset = CliSatSomeIdentifier(email = ..., )  # all parameter specific to the dataset
download(dataset,Date(2018,1),region) # only generic space/time query information 

Of course, this is just a suggestion. Thank you for this useful package.

[natgeo-wong]

Urgh sorry, yes it is most definitely mtpw2m and not mptw2m, my bad, will fix it asap!

As for abstract types, this looks like a very interesting angle to tackle it, except that I have no idea how to construct types, coming from a purely MATLAB background where I generally don't use types explicitly. Will need to think about this some more, but thanks so much for the suggestion!

[Alexander-Barth]

Types (https://docs.julialang.org/en/v1/manual/types/) are pretty neat in julia. For example:

julia> mutable struct MyDataset
          email::String
       end

julia> ds = MyDataset("foo@bar.com");

julia> import Base: download

julia> function download(ds::MyDataset,time)
          @show "downloading $time using email $(ds.email)"
          return "file.nc"
       end
download (generic function with 7 methods)

julia> download(ds,123)
"downloading $(time) using email $(ds.email)" = "downloading 123 using email foo@bar.com"
"file.nc"

So you can have multiple download function for different types.