don't try to guess encodings

lxml works best with bytes, so encode strings as utf-8 bytes Closes #76
laws-africa · Aug 30, 2023 · f9e3aff · f9e3aff
1 parent a8f50f0
commit f9e3aff
Show file tree

Hide file tree

Showing 2 changed files with 37 additions and 6 deletions.
diff --git a/cobalt/akn.py b/cobalt/akn.py
@@ -70,13 +70,8 @@ class AkomaNtosoDocument:
     source = ["cobalt", "cobalt", "https://github.com/laws-africa/cobalt"]
 
     def __init__(self, xml=None):
-        # TODO: we can do this better
-        encoding = ENCODING_RE.search(xml, 0, 200)
-        if encoding:
-            # lxml doesn't like unicode strings with an encoding element, so
-            # change to bytes
+        if isinstance(xml, str):
             xml = xml.encode('utf-8')
-
         self.parse(xml)
         self.maker = objectify.ElementMaker(annotate=False, namespace=self.namespace, nsmap=self.root.nsmap)
 

diff --git a/tests/test_structured_document.py b/tests/test_structured_document.py
@@ -194,6 +194,42 @@ def test_parser(self):
   </act>
 </akomaNtoso>""", a.document_type)
 
+    def test_unicode(self):
+        # string, no encoding
+        a = Act("""<akomaNtoso xmlns="http://docs.oasis-open.org/legaldocml/ns/akn/3.0">
+          <act>
+            <meta/>
+            <body>😀</body>
+          </act>
+        </akomaNtoso>""")
+        self.assertEqual(a.root.xpath("//a:body", namespaces={'a': a.namespace})[0].text, "😀")
+
+        # bytes, no encoding
+        a = Act("""<akomaNtoso xmlns="http://docs.oasis-open.org/legaldocml/ns/akn/3.0">
+          <act>
+            <meta/>
+            <body>😀</body>
+          </act>
+        </akomaNtoso>""".encode('utf-8'))
+        self.assertEqual(a.root.xpath("//a:body", namespaces={'a': a.namespace})[0].text, "😀")
+
+        # with encoding attribute, bytes
+        a = Act("""<?xml version="1.0" encoding="utf-8"?><akomaNtoso xmlns="http://docs.oasis-open.org/legaldocml/ns/akn/3.0">
+          <act>
+            <meta/>
+            <body>😀</body>
+          </act>
+        </akomaNtoso>""".encode('utf-8'))
+        self.assertEqual(a.root.xpath("//a:body", namespaces={'a': a.namespace})[0].text, "😀")
+
+        # with encoding string
+        Act("""<?xml version="1.0" encoding="utf-8"?><akomaNtoso xmlns="http://docs.oasis-open.org/legaldocml/ns/akn/3.0">
+          <act>
+            <meta/>
+            <body>😀</body>
+          </act>
+        </akomaNtoso>""")
+
     def test_add_number(self):
         """ When adding an FRBRnumber element to a document that doesn't already have one, it
         must come after subtype.