Merge branch 'master' of ssh://github.com/honnibal/spaCy

fabfile.py

@@ -54,10 +54,10 @@ def prebuild(build_dir='/tmp/build_spacy'):
             local('pip install --no-cache-dir -r requirements.txt')
             local('fab clean make')
             local('cp -r %s/corpora/en/wordnet corpora/en/' % spacy_dir)
-            local('cp %s/corpora/en/freqs.txt.gz corpora/en/' % spacy_dir)
             local('PYTHONPATH=`pwd` python bin/init_model.py en lang_data corpora spacy/en/data')
             local('fab test')
-            local('python setup.py sdist')
+            local('PYTHONPATH=`pwd` python -m spacy.en.download --force all')
+            local('py.test --models spacy/tests/')
 
 
 def docs():
@@ -121,9 +121,8 @@ def clean():
 
 def test():
     with virtualenv(VENV_DIR):
-        # Run each test file separately. pytest is performing poorly, not sure why
         with lcd(path.dirname(__file__)):
-            local('py.test -x tests/')
+            local('py.test -x spacy/tests')
 
 
 def train(json_dir=None, dev_loc=None, model_dir=None):

spacy/tests/test_pickle.py

@@ -5,6 +5,11 @@ import pickle
 import pytest
 import tempfile
 
+try:
+    unicode
+except NameError:
+    unicode = str
+
 @pytest.mark.models
 def test_pickle_english(EN):
     file_ = io.BytesIO()
@@ -21,7 +26,7 @@ def test_cloudpickle_to_file(EN):
     p = cloudpickle.CloudPickler(f)
     p.dump(EN)
     f.close()
-    loaded_en = cloudpickle.load(open(f.name))
+    loaded_en = cloudpickle.load(open(f.name, 'rb'))
     os.unlink(f.name)
     doc = loaded_en(unicode('test parse'))
     assert len(doc) == 2