Merge branch 'french-tokenizer-exceptions'

spacy/fr/__init__.py

@@ -19,7 +19,7 @@ class FrenchDefaults(BaseDefaults):
 
     @classmethod
     def create_tokenizer(cls, nlp=None):
-        cls.tokenizer_exceptions = TOKENIZER_EXCEPTIONS
+        cls.tokenizer_exceptions = get_tokenizer_exceptions()
         return super(FrenchDefaults, cls).create_tokenizer(nlp)
 
 

spacy/fr/language_data.py

@@ -2,10 +2,10 @@
 from __future__ import unicode_literals
 
 from .stop_words import STOP_WORDS
-from .tokenizer_exceptions import TOKENIZER_EXCEPTIONS, TOKEN_MATCH
+from .tokenizer_exceptions import get_tokenizer_exceptions, TOKEN_MATCH
 
 
 STOP_WORDS = set(STOP_WORDS)
 
 
-__all__ = ["STOP_WORDS", "TOKENIZER_EXCEPTIONS", "TOKEN_MATCH"]
+__all__ = ["STOP_WORDS", "get_tokenizer_exceptions", "TOKEN_MATCH"]

spacy/fr/tokenizer_exceptions.py

@@ -8,6 +8,7 @@ from ..language_data.tokenizer_exceptions import _URL_PATTERN
 from ..language_data.punctuation import ALPHA_LOWER
 
 from .punctuation import ELISION, HYPHENS
+from ._tokenizer_exceptions_list import BASE_EXCEPTIONS
 
 from ..symbols import *
 
@@ -16,11 +17,13 @@ import io
 import re
 
 
-def iter_exceptions():
+def get_exceptions():
-    with io.open(os.path.join(os.path.dirname(__file__), 'resources/tokenizer_exceptions'),
+    return BASE_EXCEPTIONS
-                 'rt', encoding='utf8') as f:
+
-        for line in f:
+    # with io.open(os.path.join(os.path.dirname(__file__), 'resources/tokenizer_exceptions'),
-            yield line.strip('\n')
+    #              'rt', encoding='utf8') as f:
+    #     for line in f:
+    #         yield line.strip('\n')
 
 
 def upper_first_letter(text):
@@ -142,7 +145,7 @@ def get_tokenizer_exceptions():
 
     HYPHEN = ['-', '‐']
 
-    base_exceptions = list(iter_exceptions())
+    base_exceptions = get_exceptions()
     infixes_exceptions = []
 
     for elision_char in ELISION:
@@ -214,6 +217,6 @@ REGULAR_EXP.append(_URL_PATTERN)
 
 TOKEN_MATCH = re.compile('|'.join('(?:{})'.format(m) for m in REGULAR_EXP), re.IGNORECASE).match
 
-TOKENIZER_EXCEPTIONS = get_tokenizer_exceptions()
+#TOKENIZER_EXCEPTIONS = get_tokenizer_exceptions()
 
-__all__ = ["TOKENIZER_EXCEPTIONS", "TOKEN_MATCH"]
+__all__ = ["get_tokenizer_exceptions", "TOKEN_MATCH"]

spacy/tests/regression/test_issue768.py

@@ -3,7 +3,7 @@ from __future__ import unicode_literals
 
 from ...language import Language
 from ...attrs import LANG
-from ...fr.language_data import TOKENIZER_EXCEPTIONS, STOP_WORDS
+from ...fr.language_data import get_tokenizer_exceptions, STOP_WORDS
 from ...language_data.punctuation import TOKENIZER_INFIXES, ALPHA
 
 import pytest
@@ -20,7 +20,7 @@ def fr_tokenizer_w_infix():
         class Defaults(Language.Defaults):
             lex_attr_getters = dict(Language.Defaults.lex_attr_getters)
             lex_attr_getters[LANG] = lambda text: 'fr'
-            tokenizer_exceptions = TOKENIZER_EXCEPTIONS
+            tokenizer_exceptions = get_tokenizer_exceptions()
             stop_words = STOP_WORDS
             infixes = TOKENIZER_INFIXES + [SPLIT_INFIX]
 

spacy/tests/regression/test_issue852.py

@@ -0,0 +1,12 @@
+# encoding: utf8
+from __future__ import unicode_literals
+
+import pytest
+
+
+@pytest.mark.parametrize('text', ["au-delàs", "pair-programmâmes",
+                                  "terra-formées", "σ-compacts"])
+def test_issue852(fr_tokenizer, text):
+    """Test that French tokenizer exceptions are imported correctly."""
+    tokens = fr_tokenizer(text)
+    assert len(tokens) == 1