From fbbbda195446be61eb8a3ea4930884e059c28e64 Mon Sep 17 00:00:00 2001
From: Adriane Boyd <adrianeboyd@gmail.com>
Date: Mon, 2 Aug 2021 19:07:19 +0200
Subject: [PATCH] Fix start/end chars for empty and out-of-bounds spans (#8816)

---
 spacy/tests/doc/test_span.py | 30 ++++++++++++++++++++++++++++++
 spacy/tokens/span.pyx        |  9 +++++++--
 2 files changed, 37 insertions(+), 2 deletions(-)

diff --git a/spacy/tests/doc/test_span.py b/spacy/tests/doc/test_span.py
index 6e34f2126..01b022b9d 100644
--- a/spacy/tests/doc/test_span.py
+++ b/spacy/tests/doc/test_span.py
@@ -357,6 +357,9 @@ def test_span_eq_hash(doc, doc_not_parsed):
     assert hash(doc[0:2]) != hash(doc[1:3])
     assert hash(doc[0:2]) != hash(doc_not_parsed[0:2])
 
+    # check that an out-of-bounds is not equivalent to the span of the full doc
+    assert doc[0 : len(doc)] != doc[len(doc) : len(doc) + 1]
+
 
 def test_span_boundaries(doc):
     start = 1
@@ -369,6 +372,33 @@ def test_span_boundaries(doc):
     with pytest.raises(IndexError):
         span[5]
 
+    empty_span_0 = doc[0:0]
+    assert empty_span_0.text == ""
+    assert empty_span_0.start == 0
+    assert empty_span_0.end == 0
+    assert empty_span_0.start_char == 0
+    assert empty_span_0.end_char == 0
+
+    empty_span_1 = doc[1:1]
+    assert empty_span_1.text == ""
+    assert empty_span_1.start == 1
+    assert empty_span_1.end == 1
+    assert empty_span_1.start_char == empty_span_1.end_char
+
+    oob_span_start = doc[-len(doc) - 1 : -len(doc) - 10]
+    assert oob_span_start.text == ""
+    assert oob_span_start.start == 0
+    assert oob_span_start.end == 0
+    assert oob_span_start.start_char == 0
+    assert oob_span_start.end_char == 0
+
+    oob_span_end = doc[len(doc) + 1 : len(doc) + 10]
+    assert oob_span_end.text == ""
+    assert oob_span_end.start == len(doc)
+    assert oob_span_end.end == len(doc)
+    assert oob_span_end.start_char == len(doc.text)
+    assert oob_span_end.end_char == len(doc.text)
+
 
 def test_span_lemma(doc):
     # span lemmas should have the same number of spaces as the span
diff --git a/spacy/tokens/span.pyx b/spacy/tokens/span.pyx
index 093b2a4da..48c6053c1 100644
--- a/spacy/tokens/span.pyx
+++ b/spacy/tokens/span.pyx
@@ -105,13 +105,18 @@ cdef class Span:
         if label not in doc.vocab.strings:
             raise ValueError(Errors.E084.format(label=label))
 
+        start_char = doc[start].idx if start < doc.length else len(doc.text)
+        if start == end:
+            end_char = start_char
+        else:
+            end_char = doc[end - 1].idx + len(doc[end - 1])
         self.c = SpanC(
             label=label,
             kb_id=kb_id,
             start=start,
             end=end,
-            start_char=doc[start].idx if start < doc.length else 0,
-            end_char=doc[end - 1].idx + len(doc[end - 1]) if end >= 1 else 0,
+            start_char=start_char,
+            end_char=end_char,
         )
         self._vector = vector
         self._vector_norm = vector_norm