Download and clip GEDI data

[crich011]

In order to get ground-truth data on forest height, we need to download Level 1B, Level 2A & 2B data for our regions of interest. Because of the volume and spatial extent of the data, we need to clip the region of interest from the GEDI swathes before discarding the rest of the swathe and processing the next.

[crich011]

At what point should we work on migration from R to python for this code? cc: @ShafatRahman

[ryanmcarlson]

Goals for the next week:

• Clean-up the readability of the code
• Download H5 files for at least one bbox
• Look into migrating to python??

[ShafatRahman]

I have found some python packages that might be helpful for the python migration, e.g. pyGedi to replace RGedi and h5py for handling hdf5 files. There is also an rpy package that lets you write R code within a python script, but I do not think that is the kind of migration we are looking for. Over the next few days I will be looking into pyGedi and h5py to see if those will actually work. Please let me know if you are familiar with these packages and have suggestions about how to use them or not to use them at all.

[ryanmcarlson]

I have found some python packages that might be helpful for the python migration, e.g. pyGedi to replace RGedi and h5py for handling hdf5 files. There is also an rpy package that lets you write R code within a python script, but I do not think that is the kind of migration we are looking for. Over the next few days I will be looking into pyGedi and h5py to see if those will actually work. Please let me know if you are familiar with these packages and have suggestions about how to use them or not to use them at all.

I've never used those packages before, but I can imagine that they work similarly to their R counterparts, especially if pyGEDI is distributed by the GEDI team.

[crich011]

Looking at the available information on pyGEDI, it seems like a pretty 1:1 port of RGedi. The one thing I am not really sure of and not in a position to know is whether there are any concerns about this having the issues with coordinate systems that Aron was mentioning.

All things considered, this like this is a pretty useful tool, and I am encouraged that a GEE connection is mentioned as a potential item for the v0.3. I suspect moving to python will help a lot with maintainability, and now seems as good a time as any to do that.

I think a good first step on this could be to try and take the existing R script and port it over to python. The example notebooks (see: waveforms, 2A, 2B) may or may not be useful.

[ShafatRahman]

I am not sure if pyGEDI supports filtering the data using a date range (look at the gedi_download function here). The gedi_download function is calling a web service called GEDI Finder, which is returning a response sorted by date, so we would have to write our own function to download/sort the data(?).

Here is how RGedi is handling the response from GEDI Finder.