From 68d1cdad625e6e5387912ac08f48f0d9b2f7058c Mon Sep 17 00:00:00 2001
From: Matthew Honnibal <honnibal@gmail.com>
Date: Fri, 7 Nov 2014 04:42:31 +1100
Subject: [PATCH] * When encoding POS/NER tags, accept '-' as a missing value

---
 spacy/pos_util.py | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/spacy/pos_util.py b/spacy/pos_util.py
index 039e6b15d..e5716665e 100644
--- a/spacy/pos_util.py
+++ b/spacy/pos_util.py
@@ -36,6 +36,8 @@ def read_gold(file_, tag_list, col):
     return golds
 
 def _encode_pos(tag, tag_ids, tag_list):
+    if tag == '-':
+        return 0
     if tag not in tag_ids:
         tag_ids[tag] = len(tag_list)
         tag_list.append(tag)