From 99de44d8644729f170301883d7ebc4164cd2db13 Mon Sep 17 00:00:00 2001
From: Matthew Honnibal <honnibal+gh@gmail.com>
Date: Fri, 30 Sep 2016 20:00:21 +0200
Subject: [PATCH] Changes to Doc and Token for new string store scheme

---
 spacy/tokens/doc.pyx   | 12 +++++++++---
 spacy/tokens/token.pyx | 16 +++++++++-------
 2 files changed, 18 insertions(+), 10 deletions(-)

diff --git a/spacy/tokens/doc.pyx b/spacy/tokens/doc.pyx
index 66654482e..09d8a439d 100644
--- a/spacy/tokens/doc.pyx
+++ b/spacy/tokens/doc.pyx
@@ -132,6 +132,12 @@ cdef class Doc:
                 # must be created.
                 self.push_back(
                     <const LexemeC*>self.vocab.get(self.mem, orth), has_space)
+
+    def __dealloc__(self):
+        if self.mem is not None \
+        and self.vocab is not None \
+        and self.vocab.strings is not None:
+            self.vocab.strings.remove_oov_map(self.mem)
     
     def __getitem__(self, object i):
         '''
@@ -600,14 +606,14 @@ cdef class Doc:
         if tag in self.vocab.morphology.tag_map:
             self.vocab.morphology.assign_tag(token, tag)
         else:
-            token.tag = self.vocab.strings[tag]
-        token.lemma = self.vocab.strings[lemma]
+            token.tag = self.vocab.strings.intern(tag)
+        token.lemma = self.vocab.strings.intern(lemma, mem=self.mem)
         if ent_type == 'O':
             token.ent_iob = 2
             token.ent_type = 0
         else:
             token.ent_iob = 3
-            token.ent_type = self.vocab.strings[ent_type]
+            token.ent_type = self.vocab.strings.intern(ent_type)
         # Begin by setting all the head indices to absolute token positions
         # This is easier to work with for now than the offsets
         # Before thinking of something simpler, beware the case where a dependency
diff --git a/spacy/tokens/token.pyx b/spacy/tokens/token.pyx
index 9320cb85a..50a62fb0b 100644
--- a/spacy/tokens/token.pyx
+++ b/spacy/tokens/token.pyx
@@ -1,3 +1,5 @@
+from __future__ import unicode_literals
+
 from libc.string cimport memcpy
 from cpython.mem cimport PyMem_Malloc, PyMem_Free
 # Compiler crashes on memory view coercion without this. Should report bug.
@@ -438,19 +440,19 @@ cdef class Token:
 
     property orth_:
         def __get__(self):
-            return self.vocab.strings.decode_int(self.c.lex.orth, mem=self.mem)
+            return self.vocab.strings.decode_int(self.c.lex.orth, mem=self.doc.mem)
 
     property lower_:
         def __get__(self):
-            return self.vocab.strings.decode_int(self.c.lex.lower, mem=self.mem)
+            return self.vocab.strings.decode_int(self.c.lex.lower, mem=self.doc.mem)
 
     property norm_:
         def __get__(self):
-            return self.vocab.strings.decode_int(self.c.lex.norm, mem=self.mem)
+            return self.vocab.strings.decode_int(self.c.lex.norm, mem=self.doc.mem)
 
     property shape_:
         def __get__(self):
-            return self.vocab.strings.decode_int(self.c.lex.shape, mem=self.mem)
+            return self.vocab.strings.decode_int(self.c.lex.shape, mem=self.doc.mem)
 
     property prefix_:
         def __get__(self):
@@ -462,11 +464,11 @@ cdef class Token:
 
     property lang_:
         def __get__(self):
-            return self.vocab.strings.decode_int(self.c.lex.lang, mem=self.mem)
+            return self.vocab.strings.decode_int(self.c.lex.lang, mem=self.doc.mem)
 
     property lemma_:
         def __get__(self):
-            return self.vocab.strings.decode_int(self.c.lemma, mem=self.mem)
+            return self.vocab.strings.decode_int(self.c.lemma, mem=self.doc.mem)
 
     property pos_:
         def __get__(self):
@@ -474,7 +476,7 @@ cdef class Token:
 
     property tag_:
         def __get__(self):
-            return self.vocab.strings.decode_int(self.c.tag, mem=self.mem)
+            return self.vocab.strings.decode_int(self.c.tag, mem=self.doc.mem)
 
     property dep_:
         def __get__(self):