Added a setup.py and made path to tables and models configurable

README.md

@@ -14,6 +14,10 @@ This code is written in python. To use it you will need:
 * [Keras](https://github.com/fchollet/keras) (for Semantic-Relatedness experiments only)
 * [gensim](https://radimrehurek.com/gensim/) (for vocabulary expansion when training new models)
 
+Or you can navigate to the skip-thoughts folder and run:
+
+    python setup.py install
+
 ## Getting started
 
 You will first need to download the model files and word embeddings. The embedding files (utable and btable) are quite large (>2GB) so make sure there is enough space available. The encoder vocabulary can be found in dictionary.txt.
@@ -28,12 +32,14 @@ You will first need to download the model files and word embeddings. The embeddi
 
 NOTE to Toronto users: You should be able to run the code as is from any machine, without having to download.
 
-Once these are downloaded, open skipthoughts.py and set the paths to the above files (path_to_models and path_to_tables). Now you are ready to go. Make sure to set the THEANO_FLAGS device if you want to use CPU or GPU.
+Now you are ready to go. Make sure to set the THEANO_FLAGS device if you want to use CPU or GPU.
 
 Open up IPython and run the following:
 
+    path_to_models = '/path/to/models/'
+    path_to_tables = '/path/to/tables/'
     import skipthoughts
-    model = skipthoughts.load_model()
+    model = skipthoughts.load_model(path_to_models, path_to_tables)
 
 Now suppose you have a list of sentences X, where each entry is a string that you would like to encode. To get vectors, just run the following:
 

setup.py

@@ -0,0 +1,17 @@
+#!/usr/bin/env python
+import os
+from setuptools import setup, find_packages
+
+# Utility function to read the README file for long description
+def read(fname):
+    return open(os.path.join(os.path.dirname(__file__), fname)).read()
+
+setup(name='Skip Thoughts',
+      version='1.0',
+      description='Sent2Vec encoder and training code from the paper Skip-Thought Vectors.',
+      author='Ryan Kiros',
+      url='https://github.com/ryankiros/skip-thoughts',
+      long_description=read('README.md'),
+      packages=find_packages(exclude=['contrib', 'docs', 'tests']),
+      install_requires=['theano','keras','nltk','scikit-learn','gensim','scipy']
+     )

skipthoughts.py

@@ -17,21 +17,13 @@
 
 profile = False
 
-#-----------------------------------------------------------------------------#
-# Specify model and table locations here
-#-----------------------------------------------------------------------------#
-path_to_models = '/u/rkiros/public_html/models/'
-path_to_tables = '/u/rkiros/public_html/models/'
-#-----------------------------------------------------------------------------#
-
-path_to_umodel = path_to_models + 'uni_skip.npz'
-path_to_bmodel = path_to_models + 'bi_skip.npz'
-
-
-def load_model():
+def load_model(path_to_models='./models/', path_to_tables='./models/'):
     """
     Load the model with saved tables
     """
+    path_to_umodel = path_to_models + 'uni_skip.npz'
+    path_to_bmodel = path_to_models + 'bi_skip.npz'
+
     # Load model options
     print 'Loading model parameters...'
     with open('%s.pkl'%path_to_umodel, 'rb') as f:
@@ -56,7 +48,7 @@ def load_model():
 
     # Tables
     print 'Loading tables...'
-    utable, btable = load_tables()
+    utable, btable = load_tables(path_to_tables)
 
     # Store everything we need in a dictionary
     print 'Packing up...'
@@ -71,7 +63,7 @@ def load_model():
     return model
 
 
-def load_tables():
+def load_tables(path_to_tables):
     """
     Load the tables
     """