From f7ffd94e6a0efd0fbc2488cb26a7027c0bae89c8 Mon Sep 17 00:00:00 2001
From: Matthew Honnibal <honnibal@gmail.com>
Date: Fri, 17 Apr 2015 01:40:53 +0200
Subject: [PATCH] * Add Token.conjuncts property

---
 spacy/tokens.pyx        | 22 ++++++++++++++++++++++
 tests/test_conjuncts.py | 34 ++++++++++++++++++++++++++++++++++
 2 files changed, 56 insertions(+)
 create mode 100644 tests/test_conjuncts.py

diff --git a/spacy/tokens.pyx b/spacy/tokens.pyx
index 27d99a045..4e81b8c24 100644
--- a/spacy/tokens.pyx
+++ b/spacy/tokens.pyx
@@ -11,6 +11,7 @@ from .typedefs cimport LEMMA
 from .typedefs cimport ID, ORTH, NORM, LOWER, SHAPE, PREFIX, SUFFIX, LENGTH, CLUSTER
 from .typedefs cimport POS, LEMMA, TAG, DEP
 from .parts_of_speech import UNIV_POS_NAMES
+from .parts_of_speech cimport CONJ, PUNCT
 from .lexeme cimport check_flag
 from .spans import Span
 from .structs cimport UniStr
@@ -538,6 +539,27 @@ cdef class Token:
             return Token.cinit(self.vocab, self._string,
                                self.c + self.c.head, self.i + self.c.head, self.array_len,
                                self._seq)
+        
+    property conjuncts:
+        def __get__(self):
+            """Get a list of conjoined words"""
+            cdef Token word
+            conjs = []
+            if self.c.pos != CONJ and self.c.pos != PUNCT:
+                seen_conj = False
+                for word in reversed(list(self.lefts)):
+                    if word.c.pos == CONJ:
+                        seen_conj = True
+                    elif seen_conj and word.c.pos == self.c.pos:
+                        conjs.append(word)
+            conjs.reverse()
+            conjs.append(self)
+            if seen_conj:
+                return conjs
+            elif self is not self.head and self in self.head.conjuncts:
+                return self.head.conjuncts
+            else:
+                return []
 
     property ent_type:
         def __get__(self):
diff --git a/tests/test_conjuncts.py b/tests/test_conjuncts.py
new file mode 100644
index 000000000..34643183a
--- /dev/null
+++ b/tests/test_conjuncts.py
@@ -0,0 +1,34 @@
+"""Test the Token.conjuncts property"""
+from __future__ import unicode_literals
+
+from spacy.en import English
+import pytest
+
+NLU = English()
+
+def orths(tokens):
+    return [t.orth_ for t in tokens]
+
+
+def test_simple_two():
+    tokens = NLU('I lost money and pride.')
+    pride = tokens[4]
+    assert orths(pride.conjuncts) == ['money', 'pride']
+    money = tokens[2]
+    assert orths(money.conjuncts) == ['money', 'pride']
+
+
+def test_comma_three():
+    tokens = NLU('I found my wallet, phone and keys.')
+    keys = tokens[-2]
+    assert orths(keys.conjuncts) == ['wallet', 'phone', 'keys']
+    wallet = tokens[3]
+    assert orths(wallet.conjuncts) == ['wallet', 'phone', 'keys']
+
+
+def test_and_three():
+    tokens = NLU('I found my wallet and phone and keys.')
+    keys = tokens[-2]
+    assert orths(keys.conjuncts) == ['wallet', 'phone', 'keys']
+    wallet = tokens[3]
+    assert orths(wallet.conjuncts) == ['wallet', 'phone', 'keys']