From b97d98783a211f485ac7bbd69dee292f31cd0849 Mon Sep 17 00:00:00 2001
From: Adriane Boyd <adrianeboyd@gmail.com>
Date: Wed, 2 Sep 2020 13:06:16 +0200
Subject: [PATCH] Fix Hungarian % tokenization (#6013)

---
 spacy/lang/hu/punctuation.py | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/spacy/lang/hu/punctuation.py b/spacy/lang/hu/punctuation.py
index 597f01b65..f827cd677 100644
--- a/spacy/lang/hu/punctuation.py
+++ b/spacy/lang/hu/punctuation.py
@@ -7,6 +7,7 @@ _concat_icons = CONCAT_ICONS.replace("\u00B0", "")
 
 _currency = r"\$¢£€¥฿"
 _quotes = CONCAT_QUOTES.replace("'", "")
+_units = UNITS.replace("%", "")
 
 _prefixes = (
     LIST_PUNCT
@@ -26,7 +27,7 @@ _suffixes = (
         r"(?<=[0-9])\+",
         r"(?<=°[FfCcKk])\.",
         r"(?<=[0-9])(?:[{c}])".format(c=_currency),
-        r"(?<=[0-9])(?:{u})".format(u=UNITS),
+        r"(?<=[0-9])(?:{u})".format(u=_units),
         r"(?<=[{al}{e}{q}(?:{c})])\.".format(
             al=ALPHA_LOWER, e=r"%²\-\+", q=CONCAT_QUOTES, c=_currency
         ),