The rio-tiler-mvt from cogeotiff

rio-tiler-mvt's People

Stargazers

Watchers

rio-tiler-mvt's Issues

potential use

@vincentsarago,

I'd like to know if it would be possible to use rio-tiler-mvt to extract individual pixels from a geotiff/numpy array and serve them further as MVT points.

I have a high res (30) raster dataset and would be curios if this could work .

thanks in advance

Vectorize shapes (instead of pixel)

Just wanted to see if it was possible. Answer: Not really but at least it was fun.

from rasterio._features import _shapes
from rasterio.rio.helpers import coords
from rasterio.transform import IDENTITY


cpdef rgb_to_hex(rgba):
    return '#{:02x}{:02x}{:02x}'.format(*rgba)


cpdef bytes shapes_encoder(
    data,
    mask,
    str layer_name = "layer",
    colormap = {},
):
    cdef float x, y
    cdef tuple rgba

    mvt = Tile()
    mvt_layer = Layer(mvt, layer_name.encode())

    for (p, v) in _shapes(data, mask, connectivity=4, transform=IDENTITY):
        feature = Polygon(mvt_layer)
        xs, ys = zip(*coords(p))
        feature.add_ring(len(xs))
        for (x, y) in zip(xs, ys):
            feature.set_point(int(x), int(y))

        if colormap:
            rgba = colormap.get(v, (0, 0, 0, 0))
            feature.add_property("rgba".encode(), rgb_to_hex(rgba).encode())
        else:
            feature.add_property("value".encode(), str(v).encode())
        feature.commit()

    return mvt.serialize()

problem:

polygon returned by rasterio.features._shapes are quite complex and do need better handling (interior ring, multi polygon)
Polygon are interesting the tile bounds and should not be polygon but LineString 🤷‍♂️

Recommend Projects