From 842782c128c89318a717837254fd0ee7b04e2e00 Mon Sep 17 00:00:00 2001
From: ines <ines@ines.io>
Date: Wed, 15 Mar 2017 17:33:29 +0100
Subject: [PATCH] Move fix_deprecated_glove_vectors_loading to deprecated.py

---
 spacy/deprecated.py  | 29 +++++++++++++++++++++++++++++
 spacy/en/__init__.py | 37 +++----------------------------------
 2 files changed, 32 insertions(+), 34 deletions(-)

diff --git a/spacy/deprecated.py b/spacy/deprecated.py
index d75354f9c..c127da28a 100644
--- a/spacy/deprecated.py
+++ b/spacy/deprecated.py
@@ -3,6 +3,7 @@ from sputnik.package_list import (PackageNotFoundException,
                                   CompatiblePackageNotFoundException)
 
 import sputnik
+from pathlib import Path
 from . import about
 
 
@@ -30,6 +31,7 @@ def get_package_by_name(name=None, via=None):
                            (lang.__module__))
 
 
+from . import util
 
 
 def read_lang_data(package):
@@ -79,4 +81,31 @@ def detokenize(token_rules, words): # Deprecated?
     return positions
 
 
+def fix_glove_vectors_loading(overrides):
+    """Special-case hack for loading the GloVe vectors, to support deprecated
+    <1.0 stuff. Phase this out once the data is fixed."""
 
+    if 'data_dir' in overrides and 'path' not in overrides:
+        raise ValueError("The argument 'data_dir' has been renamed to 'path'")
+    if overrides.get('path') is False:
+        return overrides
+    if overrides.get('path') in (None, True):
+        data_path = util.get_data_path()
+    else:
+        path = overrides['path']
+        if isinstance(path, basestring):
+            path = Path(path)
+        data_path = path.parent
+    vec_path = None
+    if 'add_vectors' not in overrides:
+        if 'vectors' in overrides:
+            vec_path = util.match_best_version(overrides['vectors'], None, data_path)
+            if vec_path is None:
+                return overrides
+        else:
+            vec_path = util.match_best_version('en_glove_cc_300_1m_vectors', None, data_path)
+        if vec_path is not None:
+            vec_path = vec_path / 'vocab' / 'vec.bin'
+    if vec_path is not None:
+        overrides['add_vectors'] = lambda vocab: vocab.load_vectors_from_bin_loc(vec_path)
+    return overrides
diff --git a/spacy/en/__init__.py b/spacy/en/__init__.py
index 97719e39e..186a3b173 100644
--- a/spacy/en/__init__.py
+++ b/spacy/en/__init__.py
@@ -2,15 +2,12 @@
 from __future__ import unicode_literals, print_function
 
 from os import path
-from pathlib import Path
-
-from ..util import match_best_version
-from ..util import get_data_path
 from ..language import Language
 from ..lemmatizer import Lemmatizer
 from ..vocab import Vocab
 from ..tokenizer import Tokenizer
 from ..attrs import LANG
+from ..deprecated import fix_glove_vectors_loading
 
 from .language_data import *
 
@@ -33,34 +30,6 @@ class English(Language):
 
 
     def __init__(self, **overrides):
-        # Make a special-case hack for loading the GloVe vectors, to support
-        # deprecated <1.0 stuff. Phase this out once the data is fixed.
-        overrides = _fix_deprecated_glove_vectors_loading(overrides)
+        # Special-case hack for loading the GloVe vectors, to support <1.0
+        overrides = fix_glove_vectors_loading(overrides)
         Language.__init__(self, **overrides)
-
-
-def _fix_deprecated_glove_vectors_loading(overrides):
-    if 'data_dir' in overrides and 'path' not in overrides:
-        raise ValueError("The argument 'data_dir' has been renamed to 'path'")
-    if overrides.get('path') is False:
-        return overrides
-    if overrides.get('path') in (None, True):
-        data_path = get_data_path()
-    else:
-        path = overrides['path']
-        if isinstance(path, basestring):
-            path = Path(path)
-        data_path = path.parent
-    vec_path = None
-    if 'add_vectors' not in overrides:
-        if 'vectors' in overrides:
-            vec_path = match_best_version(overrides['vectors'], None, data_path)
-            if vec_path is None:
-                return overrides
-        else:
-            vec_path = match_best_version('en_glove_cc_300_1m_vectors', None, data_path)
-        if vec_path is not None:
-            vec_path = vec_path / 'vocab' / 'vec.bin'
-    if vec_path is not None:
-        overrides['add_vectors'] = lambda vocab: vocab.load_vectors_from_bin_loc(vec_path)
-    return overrides