Modified helpers common_html_entity and replace_html_entity() to use …

…the HTML entity definitions from html.entities.html5

CHANGES

@@ -16,6 +16,9 @@ Version 3.0.0 -
   . added mark_control argument to support highlighting of control characters using
     '.' or Unicode symbols, such as "␍" and "␊".
 
+- Modified helpers common_html_entity and replace_html_entity() to use the HTML
+  entity definitions from html.entities.html5.
+
 
 Version 3.0.0rc2 -
 ------------------

pyparsing/helpers.py

@@ -1,4 +1,6 @@
 # helpers.py
+import html.entities
+
 from .core import *
 from .util import _bslash, _flatten, _escapeRegexRangeChars
 
@@ -648,10 +650,9 @@ def make_xml_tags(
     Word(alphas, alphanums + "_:").set_name("any tag")
 )
 
-
-_htmlEntityMap = dict(zip("gt lt amp nbsp quot apos".split(), "><& \"'"))
+_htmlEntityMap = {k.rstrip(";"): v for k, v in html.entities.html5.items()}
 common_html_entity = Regex(
-    "&(?P<entity>" + "|".join(_htmlEntityMap.keys()) + ");"
+    "&(?P<entity>" + "|".join(_htmlEntityMap) + ");"
 ).set_name("common HTML entity")
 
 

tests/test_unit.py

@@ -1860,6 +1860,31 @@ def testRecursiveCombine(self):
 
         self.assertParseResultsEquals(testVal, expected_list=expected)
 
+    def testHTMLEntities(self):
+        html_source = dedent("""\
+        This & that
+        2 > 1
+        0 < 1
+        Don't get excited!
+        I said "Don't get excited!"
+        Copyright © 2021
+        Dot ⟶ ˙
+        """)
+        transformer = pp.common_html_entity.add_parse_action(pp.replace_html_entity)
+        transformed = transformer.transform_string(html_source)
+        print(transformed)
+
+        expected = dedent("""\
+        This & that
+        2 > 1
+        0 < 1
+        Don't get excited!
+        I said "Don't get excited!"
+        Copyright © 2021
+        Dot ⟶ ˙
+        """)
+        self.assertEqual(expected, transformed)
+
     def testInfixNotationBasicArithEval(self):
         import ast