Matthew Perry
Matthew Perry
Implementation is a little trickier than expected, deferring this to the 0.12 milestone
@sgoodm Thanks! I'll look into the test failure and give this a review next week.
That sounds like a good idea but would change the interface significantly. I'd accept a pull request if you wanted to make it a separate function.
:+1: sounds good to me, probably should follow the numpy convention and name it `return_variance_dict`
A unit test would be ideal. Geopandas is included in the CI test runner so it should work.
One important note is that it's (practically) impossible to crash the python interpreter without calling into C code. In other words, pure python code will never crash - but C...
I can't reproduce this but, given we directly import click, it makes sense to include it as a dependency
@groutr Thanks for the tip - the dtype accumulator might be exactly what we need here - the array stays in its native dtype but the result can't overflow. @maxfenv...
I can't get the rasterio test passing. If we can't get file-based raster support, this is dead in the water. Any ideas @sgoodm ? In the mean time, I'll move...
The stats are listed in the same order as the input features. If you want to retain the attributes, you can pass `geojson_out=True` so that the output is a complete...