Added support for Open Graph arrays

extruct/_extruct.py

@@ -21,6 +21,7 @@ def extract(htmlstring,
             uniform=False,
             return_html_node=False,
             schema_context='http://schema.org',
+            with_og_array=False,
             **kwargs):
     """htmlstring: string with valid html document;
        base_url: base url of the html document
@@ -134,7 +135,7 @@ def extract(htmlstring,
         for syntax, uniform, raw, schema_context in uniform_processors:
             try:
                 if syntax == 'opengraph':
-                    output[syntax] = uniform(raw)
+                    output[syntax] = uniform(raw, with_og_array=with_og_array)
                 else:
                     output[syntax] = uniform(raw, schema_context)
             except Exception as e:

extruct/uniform.py

@@ -1,16 +1,29 @@
 from six.moves.urllib.parse import urlparse, urljoin
 
 
-def _uopengraph(extracted):
+def _uopengraph(extracted, with_og_array=False):
     out = []
     for obj in extracted:
         # In order of appearance in the page
-        properties = list(reversed(obj['properties']))
-        # Ensuring that never empty value is returned if there is a duplicated
-        # property with non empty value
-        non_empty_props = {k for k, v in properties if v and v.strip()}
-        flattened = {k: v for k, v in properties
-                     if k not in non_empty_props or (v and v.strip())}
+        properties = list(obj['properties'])
+        flattened = {}
+
+        for k, v in properties:
+            if k not in flattened.keys():
+                flattened[k] = v
+            elif v and v.strip():
+                # If og_array isn't required add first non empty value
+                if not with_og_array:
+                    if not flattened[k] or not flattened[k].strip():
+                        flattened[k] = v
+                else:
+                    if isinstance(flattened[k], list):
+                        flattened[k].append(v)
+                    elif flattened[k] and flattened[k].strip():
+                        flattened[k] = [flattened[k], v]
+                    else:
+                        flattened[k] = v
+
         t = flattened.pop('og:type', None)
         if t:
             flattened['@type'] = t

tests/test_uniform.py

@@ -27,6 +27,25 @@ def test_uopengraph(self):
         data = extruct.extract(body, syntaxes=['opengraph'], uniform=True)
         self.assertEqual(data['opengraph'], expected)
 
+    def test_uopengraph_with_og_array(self):
+        expected = [{"@context": {
+                        "og": "http://ogp.me/ns#",
+                        "fb": "http://www.facebook.com/2008/fbml",
+                        "concerts": "http://ogp.me/ns/fb/songkick-concerts#"
+                    },
+                    "fb:app_id": "308540029359",
+                    "og:site_name": "Songkick",
+                    "@type": "songkick-concerts:artist",
+                    "og:title": "Elysian Fields",
+                    "og:description": "Buy tickets for an upcoming Elysian Fields concert near you. List of all Elysian Fields tickets and tour dates for 2017.",
+                    "og:url": "http://www.songkick.com/artists/236156-elysian-fields",
+                    "og:image": [ "http://images.sk-static.com/images/media/img/col4/20100330-103600-169450.jpg",
+                                  "http://images.sk-static.com/SECONDARY_IMAGE.jpg"],
+                }]
+        body = get_testdata('songkick', 'elysianfields.html')
+        data = extruct.extract(body, syntaxes=['opengraph'], uniform=True, with_og_array=True)
+        self.assertEqual(data['opengraph'], expected)
+
     def test_uopengraph_duplicated_priorities(self):
         # Ensures that first seen property is kept when flattening
         data = _uopengraph([{'properties':
@@ -58,6 +77,36 @@ def test_uopengraph_duplicated_priorities(self):
         assert data[0]['prop_non_empty2'] == 'value!'
         assert data[0]['prop_non_empty3'] == 'value!'
 
+    def test_uopengraph_duplicated_with_og_array(self):
+        # Ensures that first seen property is kept when flattening
+        data = _uopengraph([{'properties':
+                                 [('prop_{}'.format(k), 'value_{}'.format(v))
+                                  for k in range(5)
+                                  for v in range(5)],
+                             'namespace': 'namespace'}], with_og_array=True)
+        for k in range(5):
+            assert data[0]['prop_{}'.format(k)] == ['value_0', 'value_1', 'value_2', 'value_3', 'value_4']
+
+        # Ensures that empty is not returned if a property contains any
+        # non empty value
+        data = _uopengraph([{'properties':
+                                 [('prop_empty', ' '),
+
+                                  ('prop_non_empty', ' '),
+                                  ('prop_non_empty', 'value!'),
+
+                                  ('prop_non_empty2', 'value!'),
+                                  ('prop_non_empty2', ' '),
+
+                                  ('prop_non_empty3', ' '),
+                                  ('prop_non_empty3', 'value!'),
+                                  ('prop_non_empty3', 'other value'),
+                                  ],
+                             'namespace': 'namespace'}], with_og_array=True)
+        assert data[0]['prop_empty'] == ' '
+        assert data[0]['prop_non_empty'] == 'value!'
+        assert data[0]['prop_non_empty2'] == 'value!'
+        assert data[0]['prop_non_empty3'] == ['value!', 'other value']
 
     def test_umicroformat(self):
         expected = [ { '@context': 'http://microformats.org/wiki/',