✨ StackSTACMosaicIterDataPipe to mosaic tiles into one piece

[weiji14]

An iterable-style DataPipe to mosaic tiles in an xarray.DataArray! Uses stackstac.mosaic to do the mosaic operation.

Preview at https://zen3geo--63.org.readthedocs.build/en/63/api.html#zen3geo.datapipes.StackSTACMosaic

Usage:

import pystac

from torchdata.datapipes.iter import IterableWrapper
from zen3geo.datapipes import StackSTACMosaic, StackSTACStacker

# Get list of ALOS DEM tiles to mosaic together later
item_urls = [
    "https://planetarycomputer.microsoft.com/api/stac/v1/collections/alos-dem/items/ALPSMLC30_N022E113_DSM",
    "https://planetarycomputer.microsoft.com/api/stac/v1/collections/alos-dem/items/ALPSMLC30_N022E114_DSM",
]

stac_items = [pystac.Item.from_file(href=url) for url in item_urls]
dp = IterableWrapper(iterable=[stac_items])

# Mosaic different tiles in an xarray.DataArray using DataPipe
dp_mosaic = dp.stack_stac_items().mosaic_dataarray()

# Loop or iterate over the DataPipe stream
it = iter(dp_mosaic)
dataarray = next(it)

print(dataarray.sizes)
# Frozen({'band': 1, 'y': 3600, 'x': 7200})

print(dataarray.coords)
# Coordinates:
#   * band         (band) <U4 'data'
#   * x            (x) float64 113.0 113.0 113.0 113.0 ... 115.0 115.0 115.0 115.0
#   * y            (y) float64 23.0 23.0 23.0 23.0 23.0 ... 22.0 22.0 22.0 22.0
#     proj:epsg    int64 4326
#     instruments  <U5 'prism'
#     platform     <U4 'alos'
#     gsd          int64 30
#     proj:shape   object {3600}
#     epsg         int64 4326

print(dataarray.attrs["spec"])
# RasterSpec(epsg=4326, bounds=(113.0, 22.0, 115.0, 23.0), resolutions_xy=(0.0002777777777777778, 0.0002777777777777778))

TODO:

• Initial implementation with a doctest
• Add unit test

Needed for #62