From 3fd3bc79aa7fa5c1c1ae360b49b3d2a1da6b0f36 Mon Sep 17 00:00:00 2001
From: "Yubing (Tom) Dong" <tom.tung.dyb@gmail.com>
Date: Wed, 7 Oct 2015 01:25:35 -0700
Subject: [PATCH] Refactor to remove duplicate slicing logic

---
 spacy/tokens/doc.pyx   | 11 +++--------
 spacy/tokens/spans.pyx | 27 +++++----------------------
 spacy/util.py          | 20 ++++++++++++++++++++
 3 files changed, 28 insertions(+), 30 deletions(-)

diff --git a/spacy/tokens/doc.pyx b/spacy/tokens/doc.pyx
index ce278d868..b78214ba9 100644
--- a/spacy/tokens/doc.pyx
+++ b/spacy/tokens/doc.pyx
@@ -21,6 +21,7 @@ from ..lexeme cimport Lexeme
 from .spans cimport Span
 from .token cimport Token
 from ..serialize.bits cimport BitArray
+from ..util import normalize_slice
 
 
 DEF PADDING = 5
@@ -87,14 +88,8 @@ cdef class Doc:
             token (Token):
         """
         if isinstance(i, slice):
-            if not (i.step is None or i.step == 1):
-                raise ValueError("Stepped slices not supported in Span objects."
-                                 "Try: list(doc)[start:stop:step] instead.")
-            if i.start is None:
-                i = slice(0, i.stop)
-            if i.stop is None:
-                i = slice(i.start, len(self))
-            return Span(self, i.start, i.stop, label=0)
+            start, stop = normalize_slice(len(self), i.start, i.stop, i.step)
+            return Span(self, start, stop, label=0)
 
         if i < 0:
             i = self.length + i
diff --git a/spacy/tokens/spans.pyx b/spacy/tokens/spans.pyx
index 955d24ad4..e8d2f2e59 100644
--- a/spacy/tokens/spans.pyx
+++ b/spacy/tokens/spans.pyx
@@ -9,19 +9,15 @@ from ..structs cimport TokenC, LexemeC
 from ..typedefs cimport flags_t, attr_t
 from ..attrs cimport attr_id_t
 from ..parts_of_speech cimport univ_pos_t
+from ..util import normalize_slice
 
 
 cdef class Span:
     """A slice from a Doc object."""
     def __cinit__(self, Doc tokens, int start, int end, int label=0, vector=None,
                   vector_norm=None):
-        if start < 0:
-            start = tokens.length + start
-        start = min(tokens.length, max(0, start))
-
-        if end < 0:
-            end = tokens.length + end
-        end = min(tokens.length, max(start, end))
+        if not (0 <= start <= end <= len(tokens)):
+            raise IndexError
 
         self.doc = tokens
         self.start = start
@@ -52,23 +48,10 @@ cdef class Span:
 
     def __getitem__(self, object i):
         if isinstance(i, slice):
-            start, end, step = i.start, i.stop, i.step
-            if start is None:
-               start = 0
-            elif start < 0:
-               start += len(self)
-            start = min(len(self), max(0, start))
-
-            if end is None:
-               end = len(self)
-            elif end < 0:
-               end += len(self)
-            end = min(len(self), max(start, end))
-
+            start, end = normalize_slice(len(self), i.start, i.stop, i.step)
             start += self.start
             end += self.start
-
-            return self.doc[start:end:i.step]
+            return Span(self.doc, start, end)
 
         if i < 0:
             return self.doc[self.end + i]
diff --git a/spacy/util.py b/spacy/util.py
index 9f5b4fe04..449b06399 100644
--- a/spacy/util.py
+++ b/spacy/util.py
@@ -7,6 +7,26 @@ from .attrs import TAG, HEAD, DEP, ENT_IOB, ENT_TYPE
 DATA_DIR = path.join(path.dirname(__file__), '..', 'data')
 
 
+def normalize_slice(length, start, stop, step=None):
+    if not (step is None or step == 1):
+        raise ValueError("Stepped slices not supported in Span objects."
+                         "Try: list(tokens)[start:stop:step] instead.")
+    if start is None:
+       start = 0
+    elif start < 0:
+       start += length
+    start = min(length, max(0, start))
+
+    if stop is None:
+       stop = length
+    elif stop < 0:
+       stop += length
+    stop = min(length, max(start, stop))
+
+    assert 0 <= start <= stop <= length
+    return start, stop
+
+
 def utf8open(loc, mode='r'):
     return codecs.open(loc, mode, 'utf8')