Improve logic

changedetectionio/async_update_worker.py

@@ -352,7 +352,7 @@ async def async_update_worker(worker_id, q, notification_q, app, datastore):
                 try:
                     page_title = html_tools.extract_title(data=update_handler.fetcher.content)
                     logger.info(f"UUID: {uuid} Page <title> is '{page_title}")
-                    datastore.update_watch(uuid=uuid, update_obj={'title': page_title})
+                    datastore.update_watch(uuid=uuid, update_obj={'page_title': page_title})
                 except Exception as e:
                     logger.warning(f"UUID: {uuid} Exception when extracting <title> - {str(e)}")
 

changedetectionio/html_tools.py

@@ -548,7 +548,9 @@ def extract_title(data: bytes | str, sniff_bytes: int = 2048, scan_chars: int =
 
         # Search only in the prefix
         if m := TITLE_RE.search(prefix):
-            return html.unescape(" ".join(m.group(1).split())).strip()
+            title = html.unescape(" ".join(m.group(1).split())).strip()
+            # Some safe limit
+            return title[:2000]
         return None
         
     except Exception as e:

changedetectionio/model/Watch.py

@@ -169,8 +169,8 @@ class model(watch_base):
 
     @property
     def label(self):
-        # Used for sorting
-        return self.get('title') if self.get('title') else self.get('url')
+        # Used for sorting, display, etc
+        return self.get('title') or self.get('page_title') or self.get('url')
 
     @property
     def last_changed(self):

changedetectionio/model/__init__.py

@@ -48,6 +48,7 @@ class watch_base(dict):
             'notification_screenshot': False,  # Include the latest screenshot if available and supported by the apprise URL
             'notification_title': None,
             'notification_urls': [],  # List of URLs to add to the notification Queue (Usually AppRise)
+            'page_title': None, # <title> from the page
             'paused': False,
             'previous_md5': False,
             'previous_md5_before_filters': False,  # Used for skipping changedetection entirely
@@ -121,7 +122,7 @@ class watch_base(dict):
                     }
                 },
             },
-            'title': None,
+            'title': None, # An arbitrary field that overrides 'page_title'
             'track_ldjson_price_data': None,
             'trim_text_whitespace': False,
             'remove_duplicate_lines': False,

changedetectionio/templates/_common_fields.html

@@ -70,7 +70,7 @@
                                     </tr>
                                     <tr>
                                         <td><code>{{ '{{watch_title}}' }}</code></td>
-                                        <td>The page title of the watch.</td>
+                                        <td>The page title of the watch, uses &lt;title&gt; if not set.</td>
                                     </tr>
                                     <tr>
                                         <td><code>{{ '{{watch_tag}}' }}</code></td>