FlatGeobuf.jl

A native flatgeobuf implementation in Julia. Flatgeobuf is a cloud optimized geospatial vector format.

Installation

] add FlatGeobuf

Usage

import FlatGeobuf as FGB

fgb = FGB.read("countries.fgb")
FlatGeobuffer with 179 GeometryTypeMultiPolygon features.

FlatGeobuf supports the Tables.jl interface.

using DataFrames
DataFrame(fgb)
179×3 DataFrame. Omitted printing of 1 columns
│ Row │ id     │ name                                │
│     │ String │ String                              │
├─────┼────────┼─────────────────────────────────────┤
│ 1   │ ATA    │ Antarctica                          │
│ 2   │ ATF    │ French Southern and Antarctic Lands │
│ 3   │ NAM    │ Namibia                             │
│ 4   │ BWA    │ Botswana                            │
│ 5   │ SWZ    │ Swaziland                           │
│ 6   │ ZAF    │ South Africa                        │
│ 7   │ LSO    │ Lesotho                             │
│ 8   │ MOZ    │ Mozambique                          │
│ 9   │ MWI    │ Malawi                              │
│ 10  │ ZMB    │ Zambia                              │
│ 11  │ ZWE    │ Zimbabwe                            │
│ 12  │ AGO    │ Angola                              │
│ 13  │ COD    │ Democratic Republic of the Congo    │
│ 14  │ RWA    │ Rwanda                              │

FlatGeobuffers have a spatial index that we can query with a bounding box

using Extents
ex = Extent(X=(-92.73405699999999, -92.73405699999999), Y=(32.580974999999995, 32.580974999999995))
filter!(fgb, ex)
julia> DataFrame(fgb)
2×3 DataFrame. Omitted printing of 1 columns
│ Row │ id     │ name                     │
│     │ String │ String                   │
├─────┼────────┼──────────────────────────┤
│ 1   │ MEX    │ Mexico                   │
│ 2   │ USA    │ United States of America │

TODO

• Write support

Updating the Schema

I've used https://github.com/rjkat/flatbuffers-julia to autogenerate the files found in src/schema, but had to adapt them. Good autogeneration from schema is broken as of now. Probably needs waiting on a 1.6/1.7 Julia release so https://github.com/JuliaLang/julia/pull/32658 is merged, so an official flatbuffer implementation can be made over at https://github.com/google/flatbuffers/pull/5088.