Quickly create deep quadtree structures from CSV or feather files to allow serving over HTTP or quicker access to random portions of spatial data.
Each node in the quadtree is a single feather file containing some number of rows (default 2**16) and arrow metadata indicating the existence of children.
From the command line (basic):
WORK IN PROGRESS BELOW.
python -m pip install git+https://github.com/bmschmidt/quadfeather
quadfeather-test-data 255000 # populates a csv at tmp.csv with 255000 items
quadfeather --files tmp.csv --destination tiles
From a python iterable of Arrow record batches (advanced)
- Allow splitting off of some columns.
- Explain JS integration.
- Octtree for 3-d data.
- Other partition strategies? Quadtree is easy to understand.,
- Linked-open data standards.