ITemref processed after the current node, not before.

Author: ivanprado

extruct/w3cmicrodata.py

@@ -79,18 +79,18 @@ def _extract_item(self, node, items_seen, base_url):
                 item["id"] = itemid.strip()
 
         properties = collections.defaultdict(list)
-        # start with item references
+        for name, value in self._extract_properties(
+                node, items_seen=items_seen, base_url=base_url):
+            properties[name].append(value)
+
+        # process item references
         refs = node.get('itemref', '').split()
         if refs:
             for refid in refs:
                 for name, value in self._extract_property_refs(
                         node, refid, items_seen=items_seen, base_url=base_url):
                     properties[name].append(value)
 
-        for name, value in self._extract_properties(
-                node, items_seen=items_seen, base_url=base_url):
-            properties[name].append(value)
-
         props = []
         for (name, values) in properties.items():
             if not self.strict and len(values) == 1:

tests/samples/schema.org/product-ref.json

@@ -11,8 +11,8 @@
             },
             "prop3": "REFERENCED TO INCLUDE PROPERTIES AND ALSO INDIVIDUAL PRODUCT",
             "image": [
-                "img-2.jpg",
-                "anvil_executive.jpg"
+                "anvil_executive.jpg",
+                "img-2.jpg"
             ],
             "related_products": [
                 {

tests/samples/w3c/microdata.5.3.json

@@ -2,6 +2,6 @@
  {"properties": {"a": ["1", "2"], "b": ["test"]}},
  {"properties": {"a": ["1", "2"], "b": ["test"]}},
  {"properties": {"a": ["1", "2"], "b": ["test"]}},
- {"properties": {"a": ["1", "2"], "b": ["test"]}}
+ {"properties": {"a": ["2", "1"], "b": ["test"]}}
 ]
 

tests/test_microdata.py

@@ -183,5 +183,4 @@ def test_join_none(self):
 
         mde = MicrodataExtractor()
         data = mde.extract(body)
-        print(json.dumps(data, indent=4))
         self.assertEqual(data, expected)