This runs the Metadata Map, both one-dimensional and two-dimensional versions.

It is recommended that you use a virtual environment while running this. You can set one up in './ENV' using 'virtualenv -p python3 ENV'. Activate the virtual environment using '. ENV/bin/activate'. When you are done with the virtual environment, you should type 'deactivate'.

You can install the dependencies using: 'pip install -r requirements.txt'

After installing these dependencies, you will need to go to './ENV/src/activetm/activetm/tech/sampler/' and type 'make' to make the sampler. If you receive an error regarding BLAS or LAPACK, you may need to 'dnf install lapack-devel' (on Fedora, this may differ on other Linux distros) to get the development headers.

Finally, you will need a dataset. I've been using an Amazon reviews dataset, but so long as you have something that's usable by Ankura it should be usable here. You can pickle that data into a dataset.pickle file using the 'pickle_data.py' script.