* Add test for Issue #242: Overlapping matches not well recognised.

spacy/tests/matcher/test_matcher_bugfixes.py

@@ -1,8 +1,11 @@
 import pytest
 import numpy
+import os
 
+import spacy
 from spacy.matcher import Matcher
 from spacy.attrs import ORTH, LOWER, ENT_IOB, ENT_TYPE
+from spacy.attrs import ORTH, TAG, LOWER, IS_ALPHA, FLAG63
 from spacy.symbols import DATE
 
 
@@ -31,6 +34,30 @@ def test_overlap_issue118(EN):
     assert ents[0].end == 11
 
 
+def test_overlap_issue242():
+    '''Test bug from multi-word phrases breaking text representation.'''
+
+    patterns = [
+        [{LOWER: 'food'}, {LOWER: 'safety'}],
+        [{LOWER: 'safety'}, {LOWER: 'standards'}],
+    ]
+
+    if os.environ.get('SPACY_DATA'):
+        data_dir = os.environ.get('SPACY_DATA')
+    else:
+        data_dir = None
+ 
+    nlp = spacy.en.English(data_dir=data_dir, tagger=False, parser=False, entity=False)
+
+    nlp.matcher.add('FOOD', 'FOOD', {}, patterns)
+
+    doc = nlp(u'There are different food safety standards in different countries.')
+
+    food_safety, safety_standards = doc.ents
+    assert food_safety.text == u'food safety'
+    assert safety_standards.text == u'safety standards'
+
+
 def test_overlap_reorder(EN):
     '''Test order dependence'''
     doc = EN.tokenizer(u'how many points did lebron james score against the boston celtics last night')