Change to submodule structures

[martinheidegger]

Currently all data is in the master branch, which makes the master-branch very heavy. If I want to use a data-set in pnext-io, I have to use it as submodule but that would always use all the data-sets (that could become quite many).

My suggestion would be to change the way we manage data sets and have each data-set in its own branch and the master branch would keep the references to each data-set.

Here is an example master branch & referenced branches:
https://github.com/pnext/samples/tree/master2
https://github.com/pnext/samples/tree/RandomSample
https://github.com/pnext/samples/tree/NorthCarolina

@iosefa What do you think? Good idea?

[hccampos]

This could work but we'd have to be careful if we ever want to merge or maintain one branch with everything. Or maybe going full out for multiple repos?

[martinheidegger]

The master branch references to everything?! Why would we add everything?

[iosefa]

@martinheidegger what you propose would work nicely. But, if we are worried about size, then perhaps GitHub isnt the place to store the raw .las/.laz files? Maybe we could have dat links that point to them instead? Especially if we start to accrue larger (>100mb) point clouds for visualization. Actually, right now we dont really need to store the raw files on github anyway - the RandomSample dataset can be generated locally whenever it is needed and the same is true for NorthCarolina. The downside to this is that both require pdal installed (but it can be done with docker).

[iosefa]

also, i was out sick last week so sorry for not replying sooner