Add SRTM dataset

[jenkspt]

This is a first attempt at adding the SRTM dataset

Some notes on the implementation:

• The code uses a mirror of SRTM data from https://srtm.csi.cgiar.org/wp-content/uploads/files/srtm_5x5/TIFF/

• The SRTM dataset is different from the currently implemented datasets:

  • It has 5x5 degree tiles
  • It's around 400GB (uncompressed)

To handle these constraints, getraster takes a bounds argument for selecting the dataset region to download. When called with the bounds argument, this method also returns a matrix of strings with the appropriate tile paths.

TODO: Need to propose changes so that Raster(SRTM, bounds, ...) from Rasters.jl can be implemented

[jenkspt]

Is there a reason the CI is running with Julia 1.5? is it the [raster_path1;;] syntax that's causing the error?

[rafaqz]

Thanks for the PR! This looks great.

The reason for testing on 1.5 and any errors is simply that I'm spread too thin over other packages and havent had time to work on this as much as I would like to.

Github actions previously didn't recognise just "1" so you needed the specific version. But it does now, so we should use that and always test on the latest version.

[jenkspt]

I believe 1.6.5 is now the long term release. From the downloads page https://julialang.org/downloads/

[jenkspt]

Should rasterpath and rastername also have methods with the bounds argument?

[rafaqz]

Yep, and tile_indices also has to be a keyword arg everywhere.

The second argument to getraster is always the layer name/s. If there are no layers, there probably should be no second argument.

Otherwise multiple tiles means a RasterStack

[rafaqz]

We can test on 1.6 or just 1, I dont mind. I havent tested multiple versions because CI downloads a lot of files.

[jenkspt]

Ok so I removed the resolution and defres methods, and added bounds and tile_index keyword arguments to all of the relevant functions.

Tests are passing locally for me with Julia 1.7.1 -- not exactly sure why the CI is failing.

[rafaqz]

Have you missed pushing some updates to your tests?

https://github.com/EcoJulia/RasterDataSources.jl/runs/4833591681?check_suite_focus=true#step:5:1446

[codecov[bot]]

Codecov Report

Merging #46 (46ef205) into master (f19f11f) will increase coverage by 1.58%. The diff coverage is 92.30%.

@@            Coverage Diff             @@
##           master      #46      +/-   ##
==========================================
+ Coverage   73.41%   75.00%   +1.58%     
==========================================
  Files          15       16       +1     
  Lines         425      464      +39     
==========================================
+ Hits          312      348      +36     
- Misses        113      116       +3     

Impacted Files	Coverage Δ
src/RasterDataSources.jl	100.00% <ø> (ø)
src/srtm/srtm.jl	92.30% <92.30%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update f19f11f...46ef205. Read the comment docs.

[rafaqz]

Thanks! Turns out I need to use this pretty much straight away in my own work, so looking forward to testing it out.