From 31b35e3a32261efb7662c52ffe01afa97f111b0a Mon Sep 17 00:00:00 2001
From: simon987 <me@simon987.net>
Date: Sat, 28 Aug 2021 19:44:10 -0400
Subject: [PATCH] Fix clean html (again)

---
 hexlib/text.py    | 11 +++--------
 setup.py          |  2 +-
 test/test_text.py | 20 +++++++++++++++++++-
 3 files changed, 23 insertions(+), 10 deletions(-)

diff --git a/hexlib/text.py b/hexlib/text.py
index e081299..ca10b44 100644
--- a/hexlib/text.py
+++ b/hexlib/text.py
@@ -56,16 +56,11 @@ def preprocess(text, lowercase=False, clean_html=False, strip=False, remove_punc
 
     if clean_html:
         try:
-            text = XML_ENTITY_RE.sub(" ", text)
-            text = text.replace("&", " ")
-            text = text.replace("<br>", "<br/>")
-            text = "<root>" + text + "</root>"
-
-            root = etree.fromstring(text)
+            parser = etree.XMLParser(recover=True)
+            root = etree.fromstring(text, parser)
 
             text = " ".join(get_text(root))
-        except Exception as e:
-            raise e
+        except:
             pass
 
     if remove_punctuation:
diff --git a/setup.py b/setup.py
index 037487b..7420104 100644
--- a/setup.py
+++ b/setup.py
@@ -2,7 +2,7 @@ from setuptools import setup
 
 setup(
     name="hexlib",
-    version="1.45",
+    version="1.46",
     description="Misc utility methods",
     author="simon987",
     author_email="me@simon987.net",
diff --git a/test/test_text.py b/test/test_text.py
index 11c9693..5f8ea94 100644
--- a/test/test_text.py
+++ b/test/test_text.py
@@ -165,7 +165,7 @@ class TestText(TestCase):
         self.assertEqual(cleaned, expected)
 
     def test_html_no_root(self):
-        text = "<a href=\"#p217709510\" class=\"quotelink\">&gt;&gt;217709510</a><br>Is there a servant that is against civilization and humanity?<br>Literally instant summon."
+        text = "<a href=\"#p217709510\" class=\"quotelink\">&gt;&gt;217709510</a><br>Is there a<wbr>servant that is against civilization and humanity?<br>Literally instant summon."
         
         cleaned = preprocess(
             text,
@@ -180,6 +180,24 @@ class TestText(TestCase):
         )
         
         expected = "217709510 is there a servant that is against civilization and humanity literally instant summon"
+        
+
+    def test_html_invalid_attribute(self):
+        text = '<root><iframe width="560" height="315" src=" " title="youtube video player" frameborder="0" allowfullscreen></iframe></root>'
+
+        cleaned = preprocess(
+            text,
+            clean_html=True,
+            lowercase=True,
+            remove_punctuation=True,
+            strip=True,
+            lemmatize=False,
+            fix_single_quotes=True,
+            remove_stopwords_en=False,
+            remove_urls=False
+        )
+
+        expected = ""
 
         self.assertEqual(cleaned, expected)