This repo contains code exported from a research project that uses the MovieLens 100k dataset. You can’t do much of it without the context but it can be useful as a reference for various code snippets.
The project is not endorsed by the University of Minnesota or the GroupLens Research Group.
The original README follows.
Create your local Python environment with:
virtualenv venv
./venv/bin/pip install --disable-pip-version-check -r requirements.txt
This will setup the environment as well as retrieve the data for you.
If in the future you want to add another pip package install it with:
./venv/bin/pip install <the-package>
Then make sure it’s added to the dependencies:
./venv/bin/pip freeze >| requirements.txt
./import.sh <dataset>
<dataset> can be one of ml-100k, ml-1m, or ml-10m. The script downloads
the archive, unpacks it in data-<dataset>/, and imports it in the DB.
You can only have one dataset in the DB at once; if you want to switch between multiple ones you'll need to move the DB somewhere else, import another dataset then rename the files around to switch.
If you want to reset the database, for example to change the dataset, just
remove the movies.db file and run the import again.
On OS X if you get an error when using Matplotlib use
./venv/bin/frameworkpython instead of ./venv/bin/python.
Install it with:
curl -sL https://git.io/vguGk >| ./venv/bin/frameworkpython
chmod +x ./venv/bin/frameworkpython