geojson
既存のライブラリが重厚長大なものが多いのだけど、 OpenGL や SVG のような二次元のベクター描画で簡単に済ませたい。 その方向で調査。
format
GeoJSON is a geospatial data
以下のような様式。 Feature の中に Geometry が入っている。
{
"type": "FeatureCollection",
"features": [
{
"type": "Feature",
"properties": {},
"geometry": {}
}
]
}| Geometry Object |
|---|
| Point |
| MultiPoint |
| LineString |
| MultiLineString |
| Polygon |
| MultiPolygon |
| GeometryCollection |
Point
{
"type": "Point",
"coordinates": [100.0, 0.0]
}Polygon
穴が空いている場合は、複数の頂点リストを保持する。
// No holes:
{
"type": "Polygon",
"coordinates": [
[
[100.0, 0.0],
[101.0, 0.0],
[101.0, 1.0],
[100.0, 1.0],
[100.0, 0.0]
]
]
}
// with holes:
{
"type": "Polygon",
"coordinates": [
[
[100.0, 0.0],
[101.0, 0.0],
[101.0, 1.0],
[100.0, 1.0],
[100.0, 0.0]
],
[
[100.8, 0.8],
[100.8, 0.2],
[100.2, 0.2],
[100.2, 0.8],
[100.8, 0.8]
]
]
}read
python で素直に読んでみた。
import pathlib
import json
def process_geometry(geometry: dict):
match geometry:
case {"type": "MultiPolygon", "coordinates": coordinates}:
print(f'{len(coordinates)} polygon')
for coord in coordinates:
print(f' {len(coord)} rings')
for x in coord:
print(f' {len(x)} points')
def process_feature(feature: dict):
match feature:
case {"type": "Feature", "properties": props, "geometry": geometry}:
print(props)
process_geometry(geometry)
case _:
raise NotImplementedError()
def main(path: pathlib.Path):
data = json.loads(path.read_bytes())
match data:
case {"type": "FeatureCollection", "features": features}:
for feature in features:
process_feature(feature)
if __name__ == '__main__':
main(pathlib.Path('japan.geo.json'))jpan.geo.json は、 https://github.com/dataofjapan/land です。 実行結果。
{'nam': 'Kyoto Fu', 'nam_ja': '京都府', 'id': 26}
4 polygon
1 rings
1235 points
1 rings
6 points
1 rings
8 points
1 rings
6 pointsなるほど。
GL_LINE_LOOP
単純に GL_LINE_LOOP で描画できそうとわかった。
def process_geometry(geometry: dict) -> Polygon:
match geometry:
case {"type": "Polygon", "coordinates": polygon}:
assert len(polygon) == 1
array = (float2 * len(polygon[0]))()
for i, (x, y) in enumerate(polygon[0]):
array[i] = float2(x, y)
return Polygon(array, [SubMesh(0, len(array))])
case {"type": "MultiPolygon", "coordinates": polygons}:
array = (float2 * sum(len(polygon[0]) for polygon in polygons))()
i = 0
submeshes = []
for polygon in polygons:
assert len(polygon) == 1
submeshes.append(SubMesh(i, len(polygon[0])))
for (x, y) in polygon[0]:
array[i] = float2(x, y)
i += 1
return Polygon(array, submeshes)
case _:
raise NotImplementedError()orthogonal の方で適当にビューポートを (140, 35) というような適当な経度緯度に調整してやればよさそう。
data
参考
2021
1228
python で地理情報を扱う
2022
0101