todo_11-8-2018 - Githubissues

NateMietk / modeling-human-ignition

These scripts will take large historical time series stacks of daily climate variables for the US and aggregate to monthly. Steps are unpacking NetCDf, aggregate, mask/clip, and save as GeoTiff Edit Add topics

GNU General Public License v3.0

0 stars 1 forks source link

todo_11-8-2018 #15

Closed joemcglinchy closed 5 years ago

joemcglinchy commented 5 years ago

set up amazon AMI build out clip function

see https://gis.stackexchange.com/questions/154315/splitting-raster-into-tiles-using-gdal/154400

conda environment/commands for AMI

scikit-image (+ spatial libraries)

mbjoseph commented 5 years ago

Hey @joemcglinchy - a couple of quick questions:

Did you have a particular AMI that you want to build upon?
Which spatial libraries do we want to install?

joemcglinchy commented 5 years ago

I have placed an environment YAML file in the python directory which i have found to be functional on my p3 instances. That was built upon the Amazon Deep Learning AMI. I think it would make sense to build off of that AMI, no? Specific to that environment is CUDA 9.

spatial libraries would include rasterio (dep. gdal), geopandas/fiona/shapely if we are working with geometries. scipy, scikit-image, and scikit-learn for non-spatial.

mbjoseph commented 5 years ago

Ok this is done: ami-06c3ceee7d06d0ca4