fix: Add Support for lxml >= 5.2.0 (#217)

Co-authored-by: James Addison <[email protected]>
scrapinghub · Apr 11, 2024 · 9e0179b · 9e0179b
1 parent 6053812
commit 9e0179b
Show file tree

Hide file tree

Showing 5 changed files with 28 additions and 11 deletions.
diff --git a/extruct/xmldom.py b/extruct/xmldom.py
@@ -6,15 +6,22 @@
 from xml.dom import Node
 from xml.dom.minidom import Attr, NamedNodeMap
 
-from lxml.etree import (
-    ElementBase,
-    XPath,
-    _ElementStringResult,
-    _ElementUnicodeResult,
-    tostring,
-)
+from lxml.etree import ElementBase, XPath, _ElementUnicodeResult, tostring
 from lxml.html import HtmlElementClassLookup, HTMLParser
 
+try:
+    from lxml.etree import _ElementStringResult
+except ImportError:
+
+    class _ElementStringResult(bytes):  # type: ignore[no-redef]
+        """
+        _ElementStringResult is removed in lxml >= 5.1.1,
+        so we define it here for compatibility.
+        """
+
+        def getparent(self):
+            return self._parent  # type: ignore[attr-defined]
+
 
 class DomElementUnicodeResult:
     CDATA_SECTION_NODE = Node.CDATA_SECTION_NODE

diff --git a/requirements.txt b/requirements.txt
@@ -1,6 +1,6 @@
 # project requirements, install them using following command:
 # pip install -r requirements.txt
-lxml
+lxml[html_clean]
 requests
 rdflib>=6.0.0; python_version>="3.7"
 rdflib<6.0.0; python_version<"3.7"

diff --git a/tests/samples/misc/microformat_flat_test.json b/tests/samples/misc/microformat_flat_test.json
@@ -32,7 +32,10 @@
                     "aJ Styles FastLane 2018 15 x 17 Framed Plaque w/ Ring Canvas"
                 ],
                 "photo": [
-                    "/on/demandware.static/-/Sites-main/default/dwa3227ee6/images/small/CN1148.jpg"
+                    {
+                        "alt": "aJ Styles FastLane 2018 15 x 17 Framed Plaque w/ Ring Canvas",
+                        "value": "/on/demandware.static/-/Sites-main/default/dwa3227ee6/images/small/CN1148.jpg"
+                    }
                 ]
             }
         ]

diff --git a/tests/samples/misc/microformat_test.json b/tests/samples/misc/microformat_test.json
@@ -27,7 +27,11 @@
             {
                 "properties": {
                     "photo": [
-                        "/on/demandware.static/-/Sites-main/default/dwa3227ee6/images/small/CN1148.jpg"
+                        {
+                            "alt": "aJ Styles FastLane 2018 15 x 17 Framed Plaque w/ Ring Canvas",
+                            "value": "/on/demandware.static/-/Sites-main/default/dwa3227ee6/images/small/CN1148.jpg"
+
+                        }
                     ],
                     "name": [
                         "aJ Styles FastLane 2018 15 x 17 Framed Plaque w/ Ring Canvas"

diff --git a/tests/test_uniform.py b/tests/test_uniform.py
@@ -167,7 +167,10 @@ def test_umicroformat(self):
                             "Canvas"
                         ],
                         "photo": [
-                            "/on/demandware.static/-/Sites-main/default/dwa3227ee6/images/small/CN1148.jpg"
+                            {
+                                "alt": "aJ Styles FastLane 2018 15 x 17 Framed Plaque w/ Ring Canvas",
+                                "value": "/on/demandware.static/-/Sites-main/default/dwa3227ee6/images/small/CN1148.jpg",
+                            }
                         ],
                     },
                 ],