From d06275a6b092d61ad32ca11027abb9a1cc447db2 Mon Sep 17 00:00:00 2001
From: dgtlmoon <dgtlmoon@gmail.com>
Date: Fri, 12 Jul 2024 18:43:25 +0200
Subject: [PATCH] Extract <title> as title should work on all processors

---
 .../processors/text_json_diff/processor.py             |  6 ------
 changedetectionio/update_worker.py                     | 10 ++++++++++
 2 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/changedetectionio/processors/text_json_diff/processor.py b/changedetectionio/processors/text_json_diff/processor.py
index 1dc501a2..1de5bafb 100644
--- a/changedetectionio/processors/text_json_diff/processor.py
+++ b/changedetectionio/processors/text_json_diff/processor.py
@@ -338,12 +338,6 @@ class perform_site_check(difference_detection_processor):
         if blocked:
             changed_detected = False
 
-        # Extract title as title
-        if is_html:
-            if self.datastore.data['settings']['application'].get('extract_title_as_title') or watch['extract_title_as_title']:
-                if not watch['title'] or not len(watch['title']):
-                    update_obj['title'] = html_tools.extract_element(find='title', html_content=self.fetcher.content)
-
         logger.debug(f"Watch UUID {watch.get('uuid')} content check - Previous MD5: {watch.get('previous_md5')}, Fetched MD5 {fetched_md5}")
 
         if changed_detected:
diff --git a/changedetectionio/update_worker.py b/changedetectionio/update_worker.py
index a5af5c2b..fee09c5d 100644
--- a/changedetectionio/update_worker.py
+++ b/changedetectionio/update_worker.py
@@ -505,6 +505,16 @@ class update_worker(threading.Thread):
                         if update_handler.xpath_data:
                             watch.save_xpath_data(data=update_handler.xpath_data)
 
+                        # Extract <title> as title if possible/requested.
+                        if self.datastore.data['settings']['application'].get('extract_title_as_title') or watch['extract_title_as_title']:
+                            if not watch['title'] or not len(watch['title']):
+                                try:
+                                    update_obj['title'] = html_tools.extract_element(find='title', html_content=update_handler.fetcher.content)
+                                    logger.info(f"UUID: {uuid} Extract <title> updated title to '{update_obj['title']}")
+                                except Exception as e:
+                                    logger.warning(f"UUID: {uuid} Extract <title> as watch title was enabled, but couldn't find a <title>.")
+
+                        # Now update after running everything
                         try:
                             self.datastore.update_watch(uuid=uuid, update_obj=update_obj)