Translate <form> into <p> and disable LXML's form cleanup.

html_sanitizer/sanitizer.py

@@ -112,6 +112,7 @@ def normalize_whitespace_in_text_or_tail(element):
         italic_span_to_em,
         tag_replacer('b', 'strong'),
         tag_replacer('i', 'em'),
+        tag_replacer('form', 'p'),
     ],
     'element_postprocessors': [
     ],
@@ -163,6 +164,8 @@ def sanitize(self, html):
             # information to convert spans into em/strong tags
             safe_attrs_only=False,
             inline_style=False,
+            # Do not strip all form tags; we will filter them below
+            forms=False
         )(doc)
 
         # walk the tree recursively, because we want to be able to remove
@@ -279,6 +282,7 @@ def sanitize(self, html):
             remove_unknown_tags=False,
             safe_attrs_only=True,
             add_nofollow=self.add_nofollow,
+            forms=False
         )(doc)
 
         html = lxml.html.tostring(doc, encoding='unicode')

html_sanitizer/tests.py

@@ -58,6 +58,15 @@ def test_01_sanitize(self):
                 '<p><strong>A</strong>, <strong>B</strong>'
                 ' und <strong>C</strong></p>',
             ),
+            (
+                '<p><form>Zeile 1</form></p>',
+                '<p>Zeile 1</p>',
+            ),
+            # Suboptimal, should be cleaned further
+            (
+                '<form><p>Zeile 2</p></form>',
+                '<p><p>Zeile 2</p></p>',
+            ),
         ]
 
         self.run_tests(entries)