You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It took long time (20 - 30 minutes on Ceres) to load all the srtm tiles in the TileSet class init function in tileset.py line 16 ~ 43.
Is it good to change the code:
load all the tiles parallelly, or
calculate the polygons from the file name (i.e. n32_w105_1arc_v3.bil) instead of opening each file (rasterio.open function) to get the polygons, or
load all the tiles at first time, save the object to a pickle file, and then load the pickle file later? Only re-create/save the pickle file after the tiles are changed.
In the file gcdl/src/library/datasets/srtm.py line 60
tile_paths = sorted(self.ds_path.glob('n*_1arc_v3.bil'))
The program only loads the tiles and gets the polygons in the north hemisphere.
In Ceres /project/geocdl/local_data/srtm folder, there are
10223 tiles, north hemisphere
4054 tiles, south hemisphere
Shall we load the tiles in the south hemisphere?
The text was updated successfully, but these errors were encountered: