Better spatial mean #260
Better spatial mean #260
Conversation
There was a problem hiding this comment.
Juste pour bien comprendre, segmentize: c'est que les vertices du polygone sont trop longues et on les sépare en plusieurs et simplify: c'est que les vertices sont trop courtes, alors on en combine plusieurs ensemble ?
Est-ce que le 1 en argument de segmentize c'est aussi en degré?
|
Exact! Les deux ont un argument "tolerance" dans les mêmes unités que les coordonnées des polygones. Donc des degrés pour tout nos cas. |
| call_kwargs["out_chunks"] = kwargs_copy.pop("out_chunks") | ||
|
|
||
| # Pre-emptive segmentization. Same threshold as xESMF, but there's not strong analysis behind this choice | ||
| geoms = shapely.segmentize(polygon.geometry, 1) |
There was a problem hiding this comment.
Am I right in thinking that this will be skipped if we provided a shapefile with a resolution finer than 1°?
There was a problem hiding this comment.
It won't do anything to the shape indeed.
geoms == shapely.segmentize(geoms, 1) if no geom has any vertex longer than 1.
Pull Request Checklist:
number) and pull request (:pull:number) has been added.What kind of change does this PR introduce?
region='global'tospatial_mean. It will simply be replace by a bbox region covering the whole globe.spatial_mean, methodxesmfwill automatically segmentize the polygons down to a resolution of 1°, to ensure a correct average.out_chunksavailable inregrid_datasetandspatial_meanif using xesmf >= 0.8spatial_meanwith methodxesmfand aGeoDataFrameas shape.Does this PR introduce a breaking change?
Shouldn't.
Other information:
Added a test for
region='global'.