Hello fellow searchers!
I have an index with some millions geographic coordinates, each one representing agricultural soil information. I have 2-3 millions hectares of data in a resolution of 25 points per hectare.
With each point (its a GeoPoint type), I have some keywords to help me aggregate these data. I already have queries to extract a lot of information with this. I can get bbox, centroids and other things from this aggregated cloud points.
What I need now is extract the polygon for a query result. I could do this in Python/GDAL using convex hull, but for each geographic coordinate I have several information and the data volume could be too high to load over http to my API. I can make filters, but I risk to get a wrong boundary because I could not have sure about how many points its enough to define the boundary.
The better way I could think is creating this Convex Hull with Elasticsearch as a result of an aggregation, but I don't know enough of Elasticsearch to know if it is possible and I don't know how to do this. There's a way to write a custom aggregation for it? Something in Python or Java that I can install in my server and call it from my queries?
There are already something like this?