Improve oEmbed previews

changelog.d/10819.feature

@@ -0,0 +1 @@
+Improve oEmbed previews by processing the author name, photo, and video information.

synapse/rest/media/v1/oembed.py

@@ -13,7 +13,7 @@
 #  limitations under the License.
 import logging
 import urllib.parse
-from typing import TYPE_CHECKING, Optional
+from typing import TYPE_CHECKING, List, Optional
 
 import attr
 
@@ -22,6 +22,8 @@
 from synapse.util import json_decoder
 
 if TYPE_CHECKING:
+    from lxml import etree
+
     from synapse.server import HomeServer
 
 logger = logging.getLogger(__name__)
@@ -31,7 +33,7 @@
 class OEmbedResult:
     # The Open Graph result (converted from the oEmbed result).
     open_graph_result: JsonDict
-    # Number of seconds to cache the content, according to the oEmbed response.
+    # Number of milliseconds to cache the content, according to the oEmbed response.
     #
     # This will be None if no cache-age is provided in the oEmbed response (or
     # if the oEmbed response cannot be turned into an Open Graph response).
@@ -119,10 +121,22 @@ def parse_oembed_response(self, url: str, raw_body: bytes) -> OEmbedResult:
             # Ensure the cache age is None or an int.
             cache_age = oembed.get("cache_age")
             if cache_age:
-                cache_age = int(cache_age)
+                cache_age = int(cache_age) * 1000
 
             # The results.
-            open_graph_response = {"og:title": oembed.get("title")}
+            open_graph_response = {
+                "og:url": url,
+            }
+
+            # Use either title or author's name as the title.
+            title = oembed.get("title") or oembed.get("author_name")
+            if title:
+                open_graph_response["og:title"] = title
+
+            # Use the provider name and as the site.
+            provider_name = oembed.get("provider_name")
+            if provider_name:
+                open_graph_response["og:site_name"] = provider_name
 
             # If a thumbnail exists, use it. Note that dimensions will be calculated later.
             if "thumbnail_url" in oembed:
@@ -137,6 +151,15 @@ def parse_oembed_response(self, url: str, raw_body: bytes) -> OEmbedResult:
                 # If this is a photo, use the full image, not the thumbnail.
                 open_graph_response["og:image"] = oembed["url"]
 
+            elif oembed_type == "video":
+                open_graph_response["og:type"] = "video.other"
+                calc_description_and_urls(open_graph_response, oembed["html"])
+                open_graph_response["og:video:width"] = oembed["width"]
+                open_graph_response["og:video:height"] = oembed["height"]
+
+            elif oembed_type == "link":
+                open_graph_response["og:type"] = "website"
+
             else:
                 raise RuntimeError(f"Unknown oEmbed type: {oembed_type}")
 
@@ -149,6 +172,14 @@ def parse_oembed_response(self, url: str, raw_body: bytes) -> OEmbedResult:
         return OEmbedResult(open_graph_response, cache_age)
 
 
+def _fetch_urls(tree: "etree.Element", tag_name: str) -> List[str]:
+    results = []
+    for tag in tree.xpath("//*/" + tag_name):
+        if "src" in tag.attrib:
+            results.append(tag.attrib["src"])
+    return results
+
+
 def calc_description_and_urls(open_graph_response: JsonDict, html_body: str) -> None:
     """
     Calculate description for an HTML document.
@@ -179,6 +210,16 @@ def calc_description_and_urls(open_graph_response: JsonDict, html_body: str) ->
     if tree is None:
         return
 
+    # Attempt to find interesting URLs (images, videos, embeds).
+    if "og:image" not in open_graph_response:
+        image_urls = _fetch_urls(tree, "img")
+        if image_urls:
+            open_graph_response["og:image"] = image_urls[0]
+
+    video_urls = _fetch_urls(tree, "video") + _fetch_urls(tree, "embed")
+    if video_urls:
+        open_graph_response["og:video"] = video_urls[0]
+
     from synapse.rest.media.v1.preview_url_resource import _calc_description
 
     description = _calc_description(tree)

tests/rest/media/v1/test_url_preview.py

@@ -620,7 +620,7 @@ def test_oembed_photo(self):
         self.assertIn(b"/matrixdotorg", server.data)
 
         self.assertEqual(channel.code, 200)
-        self.assertIsNone(channel.json_body["og:title"])
+        self.assertIn("og:url", channel.json_body)
         self.assertTrue(channel.json_body["og:image"].startswith("mxc://"))
         self.assertEqual(channel.json_body["og:image:height"], 1)
         self.assertEqual(channel.json_body["og:image:width"], 1)
@@ -633,6 +633,8 @@ def test_oembed_rich(self):
         result = {
             "version": "1.0",
             "type": "rich",
+            # Note that this provides the author, not the title.
+            "author_name": "Alice",
             "html": "<div>Content Preview</div>",
         }
         end_content = json.dumps(result).encode("utf-8")
@@ -660,9 +662,11 @@ def test_oembed_rich(self):
 
         self.pump()
         self.assertEqual(channel.code, 200)
+        # The JSON body should have a URL, but we don't really care what it is.
+        body = channel.json_body
+        body.pop("og:url")
         self.assertEqual(
-            channel.json_body,
-            {"og:title": None, "og:description": "Content Preview"},
+            body, {"og:title": "Alice", "og:description": "Content Preview"}
         )
 
     def test_oembed_format(self):
@@ -705,7 +709,7 @@ def test_oembed_format(self):
         self.assertIn(b"format=json", server.data)
 
         self.assertEqual(channel.code, 200)
-        self.assertEqual(
-            channel.json_body,
-            {"og:title": None, "og:description": "Content Preview"},
-        )
+        # The JSON body should have a URL, but we don't really care what it is.
+        body = channel.json_body
+        body.pop("og:url")
+        self.assertEqual(body, {"og:description": "Content Preview"})