Master project about recommender systems.

How to use

1. Reading recommender's specifications from config/recommender.json, we run our recommender is made through the runnables.py script, using the command:

   make run
   

2. Reading experiment's specifications from config/runs.json, we run a full experiment is made through the runnables.py script, using the command:

   make experiment
   

3. The running options are available using the command

   python3 runnables.py -h
   

Testing

1. Running the runtests.py script, will run the tests in tests.tests:

   make test
   

2. Running the runtests.py script, with arguments that contain the modules of the unittests, will run the tests in all the provided modules.

   python runtests.py test1 test2 test3 # ...
   

Database

We are mainly using citeulike datasets, we assume that the data files are in the directory data/, and the database configuration from config/config.json

To build the database, we run the command:

make rebuild-database

We execute this statement in MySQL console before importing the data:

set autocommit = 0;

Documentation

To generate the documentation we can use the command:

make docs

Linting

We are following flake8 code conventions, we can verify it using the command:

make lint-flake8

Authors

Authored by:

• Mostafa M. Mohamed <[email protected]>
• Omar Nada <[email protected]>
• Ibrahim Alshibani <[email protected]>

Requirements

• pep8
• python-coverage
• mysql-client
• mysql-server
• mysql-python
• json
• sklearn
• numpy
• sphinx
• scipy
• overrides
• tensorflow
• keras
• pandas
• lda2vec
• spaCy
• chainer
• python-dev

Issues and TODOs

https://github.com/mostafa-mahmoud/HyPRec/issues

License

HyPRec is licensed under the Apache License 2.0

Copyright © 2017 by Albert-Ludwigs-Universität Freiburg, Institut für Informatik