From 0c6f1f38910de791233faaeddb9f7455ffbac779 Mon Sep 17 00:00:00 2001
From: svlandeg <sofie.vanlandeghem@gmail.com>
Date: Thu, 18 Jun 2020 13:00:03 +0200
Subject: [PATCH] fix BiluoPushDown parsing entities

---
 spacy/gold/example.pyx                        | 2 +-
 spacy/syntax/ner.pyx                          | 5 ++---
 spacy/tests/regression/test_issue1501-2000.py | 3 +--
 3 files changed, 4 insertions(+), 6 deletions(-)

diff --git a/spacy/gold/example.pyx b/spacy/gold/example.pyx
index 8dfeffe98..663c8cc6d 100644
--- a/spacy/gold/example.pyx
+++ b/spacy/gold/example.pyx
@@ -117,7 +117,7 @@ cdef class Example:
                         i = j2i_multi[j]
                         if output[i] is None:
                             output[i] = gold_values[j]
-        if as_string:
+        if as_string and field not in ["ENT_IOB"]:
             output = [vocab.strings[o] if o is not None else o for o in output]
         return output
 
diff --git a/spacy/syntax/ner.pyx b/spacy/syntax/ner.pyx
index a79a784be..31f89ef88 100644
--- a/spacy/syntax/ner.pyx
+++ b/spacy/syntax/ner.pyx
@@ -72,11 +72,10 @@ cdef class BiluoPushDown(TransitionSystem):
                 actions[action][entity_type] = 1
         moves = ('M', 'B', 'I', 'L', 'U')
         for example in kwargs.get('gold_parses', []):
-            for i, ner_tag in enumerate(example.token_annotation.entities):
+            for ner_tag in example.get_aligned("ENT_TYPE", as_string=True):
                 if ner_tag != 'O' and ner_tag != '-':
-                    _, label = ner_tag.split('-', 1)
                     for action in (BEGIN, IN, LAST, UNIT):
-                        actions[action][label] += 1
+                        actions[action][ner_tag] += 1
         return actions
 
     @property
diff --git a/spacy/tests/regression/test_issue1501-2000.py b/spacy/tests/regression/test_issue1501-2000.py
index bb89be711..129c00d99 100644
--- a/spacy/tests/regression/test_issue1501-2000.py
+++ b/spacy/tests/regression/test_issue1501-2000.py
@@ -268,7 +268,6 @@ def test_issue1963(en_tokenizer):
     assert doc.tensor.shape == (3, 128)
 
 
-# TODO: fix
 @pytest.mark.parametrize("label", ["U-JOB-NAME"])
 def test_issue1967(label):
     config = {"learn_tokens": False, "min_action_freq": 30, "beam_width": 1, "beam_update_prob": 1.0}
@@ -284,7 +283,7 @@ def test_issue1967(label):
             "entities": [label]
         }
     )
-    ner.moves.get_actions(gold_parses=[example])
+    assert "JOB-NAME" in ner.moves.get_actions(gold_parses=[example])[1]
 
 
 def test_issue1971(en_vocab):