Collecting ideas from several separate discussions into a single ticket: it could be useful to have an "EDR Lite" subset of the data that is small enough to download to a laptop. This could be used as the input basis for the tutorials to enable people to run them locally (i.e. not at NERSC), using either local installation of code or a desidocker container. It could also be used for local development work by having a pre-organized subset of data to work with. This might be a literally subset of the files, or it might even do something like trim the fuji zcatalog files to match only the subset of spectra that are included.
Depending upon the size, this could be released as a tarball, or perhaps as a "run this script with a bug of wgets" to get the data.
As a starter, someone could try to make an edrlite directory at NERSC at point $DESI_ROOT to that, and see what files we'd need to copy into there for the tutorials to work.
Collecting ideas from several separate discussions into a single ticket: it could be useful to have an "EDR Lite" subset of the data that is small enough to download to a laptop. This could be used as the input basis for the tutorials to enable people to run them locally (i.e. not at NERSC), using either local installation of code or a desidocker container. It could also be used for local development work by having a pre-organized subset of data to work with. This might be a literally subset of the files, or it might even do something like trim the fuji zcatalog files to match only the subset of spectra that are included.
Depending upon the size, this could be released as a tarball, or perhaps as a "run this script with a bug of wgets" to get the data.
As a starter, someone could try to make an edrlite directory at NERSC at point $DESI_ROOT to that, and see what files we'd need to copy into there for the tutorials to work.