Add API to retrieve projected user/item biases

[jfkirk]

Add API to retrieve projected user/item biases

[jfkirk]

@ThoseGrapefruits you expressed interest via Slack -- let me know if you want to take a stab at this!

[ThoseGrapefruits]

Yep, sounds good!

[ThoseGrapefruits]

Not sure when you’re getting back from 🇸🇪👨‍👩‍👧‍👦🔨 but I have a couple intro questions already:

1. What’s the workflow for setting up the environment and running tests? I’ve been out of the Python environment for too long and all I remember is pip install -r requirements.txt and I’m not sure even that is right in this context.
2. Any style requirements? I noticed a 120-char line limit seems standard.
3. What version of Python is this using? I don't want to start using 3.4-6 features if this is meant to be in 2.x.

[jfkirk]

Hey @ThoseGrapefruits !

1. After cloning, you should be able to run the tests by stepping through

pip install -r requirements.txt
pip install -r requirements_test.txt
nosetests test/

2. 120-char line limit is correct (you can find it defined in setup.cfg). Beyond that, if Flake passes I'm probably comfortable with the style. I'm also a big fan of NumPy's commit message standards: https://docs.scipy.org/doc/numpy-1.13.0/dev/gitwash/development_workflow.html

3. Currently TensorRec is checked for compatibility with Python 2.7 and 3.6.1 -- either one should work for your development, just make sure to check both (Travis will check both on your PR, too).

Thanks!

[ThoseGrapefruits]

All sounds good.

I got it working in a virtualenv, but had to use python -m nose test/ to run tests instead. Anyways, it works, and all 58 tests passed for me. Thanks!

I'll let you know when I have other questions :)