Data

The library assumes that the data is stored as a Zarr group, adhering to the conventions encoded in the ndpyramid utility for building multi-scale pyramids, including:

• Use of Web Mercator projection
• Use of consistent pixels_per_tile across scales
• Use of coordinates that extend over full globe (recreates the map zoom level quadtree pattern)

Structure

This results in a file structure that would resemble the following for average temperature data:

/
 ├── .zmetadata
 ├── 0
 │   ├── tavg
 │       └── 0.0
 ├── 1
 │   ├── tavg
 │       └── 0.0
 │       └── 0.1
 │       └── 1.0
 │       └── 1.1
 ├── 2
...

Note the quadrupling of the number of chunks as zoom level increases. This is a trait of the map zoom level quadtree mentioned earlier.

Multi-scale schema

See an example .zmetadata['.zattrs'] value from a pyramid created for a demo:

{
  "multiscales": [
    {
      "datasets": [
        {
          "path": "0",
          "pixels_per_tile": 128
        },
        {
          "path": "1",
          "pixels_per_tile": 128
        }
        ...
      ],
      "metadata": {
        "args": [],
        "method": "pyramid_reproject",
        "version": "0.0.post64"
      },
      "type": "reduce"
    }
  ]
}