Seems like there's still a format issue even when all three layers exist. No errors are generated when the map is loaded. I'm guessing it's something to how fiona is reading the new shapefiles. Does it expect it it to have a particular hierarchy?

http://mako.iri.columbia.edu:16526/fbfmaproom/

I found three distinct problems:

1. The zip file doesn't have the structure that fiona expects. I'm not sure how it should be structured, so for now I've unpacked the zip.
2. The Madagascar shapes are in a projected coordinate system, with units of meters, whereas the application expects a geographic coordinate system, with units of degrees. The Djibouti shapes worked because they use a geographic coordiate system.
3. A bug in the application code caused the error mesage to be silently discarded instead of being printed on the console.

I've pushed changes to this branch to fix 1 and 3. @nitinmagima please pull my changes, and then see if you can figure out how to reproject/deproject the shapes.

Also please share your script for constructing and deprojecting the shapes, because eventually I'd like to put that logic in the application. If the application can handle shapes in whatever form we find them, rather than requiring pre-processing, it will make it easier to add new shapes.

@mm5330 @nitinmagima what is going on? Github says Max just pushed two commits authored by Nitin that duplicate work Max did on a different branch.

There should be nothing about Ethiopia on this branch. This branch is for Madagascar changes.

@mm5330 @nitinmagima what is going on? Github says Max just pushed two commits authored by Nitin that duplicate work Max did on a different branch.

There should be nothing about Ethiopia on this branch. This branch is for Madagascar changes.

Oh weird, this wasn't me. Not sure how that happened...

I've added Nitin's script to this repo, and modified it to remove the holes that dissolve left in the shapes. It seems to be working now:

http://mako.iri.columbia.edu:9998/fbfmaproom/madagascar

A couple things to note:

1. The shapes are quite slow to load even for me, located in the US. It remains to be seen how usable this is for people in Madagascar.
2. We changed the outer boundary to include the entire country (used to only include the four southernmost regions), but we haven't expanded the obs data to cover the entire country, so you get errors if you click too far north. @dantesab9 if you want to fix this, you can do so by modifying the Y (latitude) bounds in the Ingrid URLs (which now live in the config file) and rerunning fbf-update-data.py. Alternatively, since we're already modifying the shapefile, it would be easy enough to change the script to exclude regions to the north that aren't of interest. That would probably also help with the speed.

@dantesab9 let me know what you want to do.

BTW I rebased this branch to get rid of the weird repeated commits from Max. If you pulled it before and you try to pull again, it will get complicated. I don't think anyone needs to do that, but if you do, I can help.

Hi Aaron, thank for this update too, the map look very good! About the obs data coverage, I think it is best to extend to the full country since we will use some districts that are not in the prior "national" polygon and also we may need to provide data for more northern districts that are part of the other, newer, red cross project. I'll take a look at the latitude question and also start working in updating the polygon IDs in the markdowns.

@dantesab9 are we making progress towards merging this? I just referred Max to it as an example to follow for Ethiopia, and I presume others will follow eventually. It would be good to get it into the main branch so it doesn't get out of sync with the rest of the code.