argentina_buenos_aires CSVs

[yolile]

Create a new spider using CSV files instead of JSON https://data.buenosaires.gob.ar/dataset/buenos-aires-compras

[yolile]

Note that this one is different from other existing spiders because we need to first download all the CSVs and only then unflatten the folder where they are saved

[yolile]

I've created a PR, but as mentioned there:

But I don't think we can merge this because the CSV files are large, and the flatten tool consumed>60% of the available memory on the server when I tried it. I've killed the spider before it finished because the memory was getting higher and higher.

So, I guess one option could be to split the CSV files before flattening them or something similar, but I'm not sure if it worth the effort