degauss-org / nlcd

add landcover variables to your data
GNU General Public License v3.0
0 stars 0 forks source link

Workaround for Pulling fst Files? #7

Closed osageorange closed 1 year ago

osageorange commented 1 year ago

I'm attempting to use this tool in a highly secure compute environment with extremely limited network access, which appears to prevent pulling the fst files from the repository in the following lines (59-60) from https://github.com/geomarker-io/addNlcdData/blob/master/R/nlcd_data_functions.R

s3_uri <- glue::glue("s3://geomarker/nlcd/nlcd_fst/nlcd_chunk_{nlcd_chunk}.fst")
nlcd_file <- s3::s3_get(s3_uri, public = T, quiet = [TRUE](https://rdrr.io/r/base/logical.html))

Is there a workaround available? Potentially including the fst files within the container?

erikarasnick commented 1 year ago

The fst files total about 48GB, which is too large to include in the container. One solution could be to download all the fst files and storing them in the secure environment before running the container. If they are stored at a specific file path within the environment, the container will recognize that the files are already there and would skip the download.

osageorange commented 1 year ago

Ok thanks, I will work with the system admins to find a way to download the files to ./nlcd_fst/ before running the container.

erikarasnick commented 1 year ago

Awesome! I think the path will have to be ./s3_downloads/geomarker/nlcd/nlcd_fst for the container to recognize. Kind of long, but an artifact of how we do automated downloads from s3.

osageorange commented 1 year ago

Thank you, that's helpful!