Add an admin route to get all the media in a room

This is intended to be used by administrators to monitor the media that is passing through their server, if they wish.

Signed-off-by: Travis Ralston <travpc@gmail.com>

1 files changed, 75 insertions, 56 deletions

diff --git a/synapse/storage/room.py b/synapse/storage/room.py
index 23688430b7..cd6899a4b5 100644
--- a/synapse/storage/room.py
+++ b/synapse/storage/room.py
@@ -533,73 +533,92 @@ class RoomStore(SQLBaseStore):
         )
         self.is_room_blocked.invalidate((room_id,))
 
+    def get_media_mxcs_in_room(self, room_id):
+        def _get_media_ids_in_room(txn):
+            local_media_ids, remote_media_ids = self._get_media_ids_in_room(txn, room_id)
+            local_media_mxcs = []
+            remote_media_mxcs = []
+
+            # Convert the IDs to MXC URIs
+            for media_id in local_media_ids:
+                local_media_mxcs.append("mxc://%s/%s" % (self.hostname, media_id))
+            for hostname, media_id in remote_media_ids:
+                remote_media_mxcs.append("mxc://%s/%s" % (hostname, media_id))
+
+            return local_media_mxcs, remote_media_mxcs
+        return self.runInteraction("get_media_ids_in_room", _get_media_ids_in_room)
+
     def quarantine_media_ids_in_room(self, room_id, quarantined_by):
         """For a room loops through all events with media and quarantines
         the associated media
         """
-        def _get_media_ids_in_room(txn):
-            mxc_re = re.compile("^mxc://([^/]+)/([^/#?]+)")
+        def _quarantine_media_in_room(txn):
+            local_media_mxcs, remote_media_mxcs = self._get_media_ids_in_room(txn, room_id)
+            total_media_quarantined = 0
 
-            next_token = self.get_current_events_token() + 1
+            # Now update all the tables to set the quarantined_by flag
 
-            total_media_quarantined = 0
+            txn.executemany("""
+                UPDATE local_media_repository
+                SET quarantined_by = ?
+                WHERE media_id = ?
+            """, ((quarantined_by, media_id) for media_id in local_media_mxcs))
 
-            while next_token:
-                sql = """
-                    SELECT stream_ordering, content FROM events
-                    WHERE room_id = ?
-                        AND stream_ordering < ?
-                        AND contains_url = ? AND outlier = ?
-                    ORDER BY stream_ordering DESC
-                    LIMIT ?
+            txn.executemany(
                 """
-                txn.execute(sql, (room_id, next_token, True, False, 100))
-
-                next_token = None
-                local_media_mxcs = []
-                remote_media_mxcs = []
-                for stream_ordering, content_json in txn:
-                    next_token = stream_ordering
-                    content = json.loads(content_json)
-
-                    content_url = content.get("url")
-                    thumbnail_url = content.get("info", {}).get("thumbnail_url")
-
-                    for url in (content_url, thumbnail_url):
-                        if not url:
-                            continue
-                        matches = mxc_re.match(url)
-                        if matches:
-                            hostname = matches.group(1)
-                            media_id = matches.group(2)
-                            if hostname == self.hostname:
-                                local_media_mxcs.append(media_id)
-                            else:
-                                remote_media_mxcs.append((hostname, media_id))
-
-                # Now update all the tables to set the quarantined_by flag
-
-                txn.executemany("""
-                    UPDATE local_media_repository
+                    UPDATE remote_media_cache
                     SET quarantined_by = ?
-                    WHERE media_id = ?
-                """, ((quarantined_by, media_id) for media_id in local_media_mxcs))
-
-                txn.executemany(
-                    """
-                        UPDATE remote_media_cache
-                        SET quarantined_by = ?
-                        WHERE media_origin AND media_id = ?
-                    """,
-                    (
-                        (quarantined_by, origin, media_id)
-                        for origin, media_id in remote_media_mxcs
-                    )
+                    WHERE media_origin AND media_id = ?
+                """,
+                (
+                    (quarantined_by, origin, media_id)
+                    for origin, media_id in remote_media_mxcs
                 )
+            )
 
-                total_media_quarantined += len(local_media_mxcs)
-                total_media_quarantined += len(remote_media_mxcs)
+            total_media_quarantined += len(local_media_mxcs)
+            total_media_quarantined += len(remote_media_mxcs)
 
             return total_media_quarantined
 
-        return self.runInteraction("get_media_ids_in_room", _get_media_ids_in_room)
+        return self.runInteraction("quarantine_media_in_room", _quarantine_media_in_room)
+
+    def _get_media_ids_in_room(self, txn, room_id):
+        mxc_re = re.compile("^mxc://([^/]+)/([^/#?]+)")
+
+        next_token = self.get_current_events_token() + 1
+        local_media_mxcs = []
+        remote_media_mxcs = []
+
+        while next_token:
+            sql = """
+                SELECT stream_ordering, content FROM events
+                WHERE room_id = ?
+                    AND stream_ordering < ?
+                    AND contains_url = ? AND outlier = ?
+                ORDER BY stream_ordering DESC
+                LIMIT ?
+            """
+            txn.execute(sql, (room_id, next_token, True, False, 100))
+
+            next_token = None
+            for stream_ordering, content_json in txn:
+                next_token = stream_ordering
+                content = json.loads(content_json)
+
+                content_url = content.get("url")
+                thumbnail_url = content.get("info", {}).get("thumbnail_url")
+
+                for url in (content_url, thumbnail_url):
+                    if not url:
+                        continue
+                    matches = mxc_re.match(url)
+                    if matches:
+                        hostname = matches.group(1)
+                        media_id = matches.group(2)
+                        if hostname == self.hostname:
+                            local_media_mxcs.append(media_id)
+                        else:
+                            remote_media_mxcs.append((hostname, media_id))
+
+        return local_media_mxcs, remote_media_mxcs