Open sgibson91 opened 5 years ago
This doesn't currently work due to issue #12 . Will upgrade to an environment.yml
file.
Moving numpy
to the dependencies
block should solve the issue of gensim
pulling a different numpy version (not fully tested as there's no data yet!). Something @LouiseABowler solved with the --no-cache-dir
argument to pip install
.
preprocessing.py
is now erroring as the data is not available. I would usually put the command line examples given in the README into a postBuild
file, but one dataset is 1.5GB which is not best practice for Binder.
Another method would be a python script the user runs before preprocessing.py
to get the data.
Upside of not using postBuild
: we're kind to the Binder team
Downside of not using postBuild
: the 1.5 GB file can take a long time to download!
We're now facing a RAM issue due to all of the data and model downloading that makes it not compatible with mybinder.org. See https://github.com/alan-turing-institute/the-turing-way/issues/401.
Added a postBuild script to run on Hub23.
@LouiseABowler This should all work now! 🎉
@sgibson91, has the IP address of the BinderHub changed? I was doing a last check before merging this PR and noticed that I'm not even reaching the Binder spinner page.
Nope. IT didn't give me credits in time and the subscription has been disabled so neither this Hub nor Hub23 are available.
Ah, that's inconvenient but at least you know what the problem is! I'll leave this PR open until the BinderHub is up and running again; no rush but could you let me know when it's back up? Thanks!
Will do
Add
runtime.txt
to pin Python version to 2.7 so that the repo will run on Binder