Merge branch '2877-update-checked-time' into only-call-send_content_changed_notification-on-change

changedetectionio/__init__.py

@@ -2,7 +2,7 @@
 
 # Read more https://github.com/dgtlmoon/changedetection.io/wiki
 
-__version__ = '0.48.04'
+__version__ = '0.48.05'
 
 from changedetectionio.strtobool import strtobool
 from json.decoder import JSONDecodeError

changedetectionio/api/api_v1.py

@@ -76,6 +76,7 @@ class Watch(Resource):
         # Return without history, get that via another API call
         # Properties are not returned as a JSON, so add the required props manually
         watch['history_n'] = watch.history_n
+        # attr .last_changed will check for the last written text snapshot on change
         watch['last_changed'] = watch.last_changed
         watch['viewed'] = watch.viewed
         return watch

changedetectionio/apprise_plugin/__init__.py

@@ -14,7 +14,7 @@ def apprise_custom_api_call_wrapper(body, title, notify_type, *args, **kwargs):
     import requests
     import json
     from urllib.parse import unquote_plus
-    from apprise.utils import parse_url as apprise_parse_url
+    from apprise.utils.parse import parse_url as apprise_parse_url
     from apprise import URLBase
 
     url = kwargs['meta'].get('url')

changedetectionio/model/Watch.py

@@ -247,37 +247,32 @@ class model(watch_base):
         bump = self.history
         return self.__newest_history_key
 
-    # Given an arbitrary timestamp, find the closest next key
-    # For example, last_viewed = 1000 so it should return the next 1001 timestamp
-    #
-    # used for the [diff] button so it can preset a smarter from_version
+    # Given an arbitrary timestamp, find the best history key for the [diff] button so it can preset a smarter from_version
     @property
-    def get_next_snapshot_key_to_last_viewed(self):
+    def get_from_version_based_on_last_viewed(self):
 
         """Unfortunately for now timestamp is stored as string key"""
         keys = list(self.history.keys())
         if not keys:
             return None
+        if len(keys) == 1:
+            return keys[0]
 
         last_viewed = int(self.get('last_viewed'))
-        prev_k = keys[0]
         sorted_keys = sorted(keys, key=lambda x: int(x))
         sorted_keys.reverse()
 
-        # When the 'last viewed' timestamp is greater than the newest snapshot, return second last
-        if last_viewed > int(sorted_keys[0]):
+        # When the 'last viewed' timestamp is greater than or equal the newest snapshot, return second newest
+        if last_viewed >= int(sorted_keys[0]):
             return sorted_keys[1]
+        
+        # When the 'last viewed' timestamp is between snapshots, return the older snapshot
+        for newer, older in list(zip(sorted_keys[0:], sorted_keys[1:])):
+            if last_viewed < int(newer) and last_viewed >= int(older):
+                return older
 
-        for k in sorted_keys:
-            if int(k) < last_viewed:
-                if prev_k == sorted_keys[0]:
-                    # Return the second last one so we dont recommend the same version compares itself
-                    return sorted_keys[1]
-
-                return prev_k
-            prev_k = k
-
-        return keys[0]
+        # When the 'last viewed' timestamp is less than the oldest snapshot, return oldest
+        return sorted_keys[-1]
 
     def get_history_snapshot(self, timestamp):
         import brotli

changedetectionio/templates/watch-overview.html

@@ -191,7 +191,7 @@
                     {% if watch.history_n >= 2 %}
 
                         {%  if is_unviewed %}
-                           <a href="{{ url_for('diff_history_page', uuid=watch.uuid, from_version=watch.get_next_snapshot_key_to_last_viewed) }}" target="{{watch.uuid}}" class="pure-button pure-button-primary diff-link">History</a>
+                           <a href="{{ url_for('diff_history_page', uuid=watch.uuid, from_version=watch.get_from_version_based_on_last_viewed) }}" target="{{watch.uuid}}" class="pure-button pure-button-primary diff-link">History</a>
                         {% else %}
                            <a href="{{ url_for('diff_history_page', uuid=watch.uuid)}}" target="{{watch.uuid}}" class="pure-button pure-button-primary diff-link">History</a>
                         {% endif %}

changedetectionio/tests/test_api.py

@@ -44,7 +44,6 @@ def set_modified_response():
 
     return None
 
-
 def is_valid_uuid(val):
     try:
         uuid.UUID(str(val))
@@ -56,8 +55,9 @@ def is_valid_uuid(val):
 def test_setup(client, live_server, measure_memory_usage):
     live_server_setup(live_server)
 
+
 def test_api_simple(client, live_server, measure_memory_usage):
-    #live_server_setup(live_server)
+#    live_server_setup(live_server)
 
     api_key = extract_api_key_from_UI(client)
 
@@ -129,6 +129,9 @@ def test_api_simple(client, live_server, measure_memory_usage):
     assert after_recheck_info['last_checked'] != before_recheck_info['last_checked']
     assert after_recheck_info['last_changed'] != 0
 
+    # #2877 When run in a slow fetcher like playwright etc
+    assert after_recheck_info['last_changed'] ==  after_recheck_info['last_checked']
+
     # Check history index list
     res = client.get(
         url_for("watchhistory", uuid=watch_uuid),

changedetectionio/tests/test_backend.py

@@ -2,7 +2,6 @@
 
 import time
 from flask import url_for
-from urllib.request import urlopen
 from .util import set_original_response, set_modified_response, live_server_setup, wait_for_all_checks, extract_rss_token_from_UI, \
     extract_UUID_from_client
 

changedetectionio/tests/test_nonrenderable_pages.py

@@ -48,7 +48,7 @@ def test_check_basic_change_detection_functionality(client, live_server, measure
     #####################
     client.post(
         url_for("settings_page"),
-        data={"application-empty_pages_are_a_change": "",
+        data={"application-empty_pages_are_a_change": "", # default, OFF, they are NOT a change
               "requests-time_between_check-minutes": 180,
               'application-fetch_backend': "html_requests"},
         follow_redirects=True
@@ -66,6 +66,14 @@ def test_check_basic_change_detection_functionality(client, live_server, measure
     res = client.get(url_for("index"))
     assert b'unviewed' not in res.data
 
+    uuid = next(iter(live_server.app.config['DATASTORE'].data['watching']))
+    watch = live_server.app.config['DATASTORE'].data['watching'][uuid]
+
+    assert watch.last_changed == 0
+    assert watch['last_checked'] != 0
+
+
+
 
     # ok now do the opposite
 
@@ -92,6 +100,10 @@ def test_check_basic_change_detection_functionality(client, live_server, measure
     # A totally zero byte (#2528) response should also not trigger an error
     set_zero_byte_response()
     client.get(url_for("form_watch_checknow"), follow_redirects=True)
+
+    # 2877
+    assert watch.last_changed == watch['last_checked']
+
     wait_for_all_checks(client)
     res = client.get(url_for("index"))
     assert b'unviewed' in res.data # A change should have registered because empty_pages_are_a_change is ON

changedetectionio/tests/test_security.py

@@ -69,22 +69,29 @@ def _runner_test_various_file_slash(client, file_uri):
     wait_for_all_checks(client)
     res = client.get(url_for("index"))
 
+    substrings = [b"URLs with hostname components are not permitted", b"No connection adapters were found for"]
+
+
     # If it is enabled at test time
     if strtobool(os.getenv('ALLOW_FILE_URI', 'false')):
-        # So it should permit it, but it should fall back to the 'requests' library giving an error
-        # (but means it gets passed to playwright etc)
-        assert b"URLs with hostname components are not permitted" in res.data
-        assert b"_runner_test_various_file_slash" in res.data # Can read this file OK
-    else:
-        # Default should be here
-        assert b'file:// type access is denied for security reasons.' in res.data
+        if file_uri.startswith('file:///'):
+            # This one should be the full qualified path to the file and should get the contents of this file
+            res = client.get(
+                url_for("preview_page", uuid="first"),
+                follow_redirects=True
+            )
+            assert b'_runner_test_various_file_slash' in res.data
+        else:
+            # This will give some error from requests or if it went to chrome, will give some other error :-)
+            assert any(s in res.data for s in substrings)
 
     res = client.get(url_for("form_delete", uuid="all"), follow_redirects=True)
     assert b'Deleted' in res.data
 
 def test_file_slash_access(client, live_server, measure_memory_usage):
     #live_server_setup(live_server)
-    # file: is permitted by default, but it will be caught by ALLOW_FILE_URI
+
+    # file: is NOT permitted by default, so it will be caught by ALLOW_FILE_URI check
 
     test_file_path = os.path.abspath(__file__)
     _runner_test_various_file_slash(client, file_uri=f"file://{test_file_path}")

changedetectionio/tests/unit/test_watch_model.py

@@ -16,7 +16,6 @@ class TestDiffBuilder(unittest.TestCase):
         watch = Watch.model(datastore_path='/tmp', default={})
         watch.ensure_data_dir_exists()
 
-        watch['last_viewed'] = 110
 
         # Contents from the browser are always returned from the browser/requests/etc as str, str is basically UTF-16 in python
         watch.save_history_text(contents="hello world", timestamp=100, snapshot_id=str(uuid_builder.uuid4()))
@@ -25,31 +24,42 @@ class TestDiffBuilder(unittest.TestCase):
         watch.save_history_text(contents="hello world", timestamp=112, snapshot_id=str(uuid_builder.uuid4()))
         watch.save_history_text(contents="hello world", timestamp=115, snapshot_id=str(uuid_builder.uuid4()))
         watch.save_history_text(contents="hello world", timestamp=117, snapshot_id=str(uuid_builder.uuid4()))
+    
+        p = watch.get_from_version_based_on_last_viewed
+        assert p == "100", "Correct 'last viewed' timestamp was detected"
 
-        p = watch.get_next_snapshot_key_to_last_viewed
-        assert p == "112", "Correct last-viewed timestamp was detected"
+        watch['last_viewed'] = 110
+        p = watch.get_from_version_based_on_last_viewed
+        assert p == "109", "Correct 'last viewed' timestamp was detected"
 
-        # When there is only one step of difference from the end of the list, it should return second-last change
         watch['last_viewed'] = 116
-        p = watch.get_next_snapshot_key_to_last_viewed
-        assert p == "115", "Correct 'second last' last-viewed timestamp was detected when using the last timestamp"
+        p = watch.get_from_version_based_on_last_viewed
+        assert p == "115", "Correct 'last viewed' timestamp was detected"
 
         watch['last_viewed'] = 99
-        p = watch.get_next_snapshot_key_to_last_viewed
-        assert p == "100"
+        p = watch.get_from_version_based_on_last_viewed
+        assert p == "100", "When the 'last viewed' timestamp is less than the oldest snapshot, return oldest"
 
         watch['last_viewed'] = 200
-        p = watch.get_next_snapshot_key_to_last_viewed
-        assert p == "115", "When the 'last viewed' timestamp is greater than the newest snapshot, return second last "
+        p = watch.get_from_version_based_on_last_viewed
+        assert p == "115", "When the 'last viewed' timestamp is greater than the newest snapshot, return second newest"
 
         watch['last_viewed'] = 109
-        p = watch.get_next_snapshot_key_to_last_viewed
+        p = watch.get_from_version_based_on_last_viewed
         assert p == "109", "Correct when its the same time"
 
         # new empty one
         watch = Watch.model(datastore_path='/tmp', default={})
-        p = watch.get_next_snapshot_key_to_last_viewed
+        p = watch.get_from_version_based_on_last_viewed
         assert p == None, "None when no history available"
 
+        watch.save_history_text(contents="hello world", timestamp=100, snapshot_id=str(uuid_builder.uuid4()))
+        p = watch.get_from_version_based_on_last_viewed
+        assert p == "100", "Correct with only one history snapshot"
+
+        watch['last_viewed'] = 200
+        p = watch.get_from_version_based_on_last_viewed
+        assert p == "100", "Correct with only one history snapshot"
+
 if __name__ == '__main__':
     unittest.main()

changedetectionio/tests/util.py

@@ -76,6 +76,14 @@ def set_more_modified_response():
     return None
 
 
+def set_empty_text_response():
+    test_return_data = """<html><body></body></html>"""
+
+    with open("test-datastore/endpoint-content.txt", "w") as f:
+        f.write(test_return_data)
+
+    return None
+
 def wait_for_notification_endpoint_output():
     '''Apprise can take a few seconds to fire'''
     #@todo - could check the apprise object directly instead of looking for this file

changedetectionio/update_worker.py

@@ -243,7 +243,6 @@ class update_worker(threading.Thread):
                 os.unlink(full_path)
 
     def run(self):
-        now = time.time()
         
         while not self.app.config.exit.is_set():
             update_handler = None
@@ -254,6 +253,7 @@ class update_worker(threading.Thread):
                 pass
 
             else:
+                fetch_start_time = time.time()
                 uuid = queued_item_data.item.get('uuid')
                 self.current_uuid = uuid
                 if uuid in list(self.datastore.data['watching'].keys()) and self.datastore.data['watching'][uuid].get('url'):
@@ -268,7 +268,6 @@ class update_worker(threading.Thread):
                     watch = self.datastore.data['watching'].get(uuid)
 
                     logger.info(f"Processing watch UUID {uuid} Priority {queued_item_data.priority} URL {watch['url']}")
-                    now = time.time()
 
                     try:
                         # Processor is what we are using for detecting the "Change"
@@ -288,6 +287,10 @@ class update_worker(threading.Thread):
 
                         update_handler.call_browser()
 
+                        # In reality, the actual time of when the change was detected could be a few seconds after this
+                        # For example it should include when the page stopped rendering if using a playwright/chrome type fetch
+                        fetch_start_time = time.time()
+
                         changed_detected, update_obj, contents = update_handler.run_changedetection(watch=watch)
 
                         # Re #342
@@ -512,7 +515,7 @@ class update_worker(threading.Thread):
 
                     if not self.datastore.data['watching'].get(uuid):
                         continue
-                    #
+
                     # Different exceptions mean that we may or may not want to bump the snapshot, trigger notifications etc
                     if process_changedetection_results:
 
@@ -525,8 +528,6 @@ class update_worker(threading.Thread):
                                 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
-                        timestamp = round(time.time())
                         try:
                             self.datastore.update_watch(uuid=uuid, update_obj=update_obj)
 
@@ -542,25 +543,28 @@ class update_worker(threading.Thread):
 
                                 # Small hack so that we sleep just enough to allow 1 second  between history snapshots
                                 # this is because history.txt indexes/keys snapshots by epoch seconds and we dont want dupe keys
-
-                                if watch.newest_history_key and int(timestamp) == int(watch.newest_history_key):
+                                # @also - the keys are one per second at the most (for now)
+                                if watch.newest_history_key and int(fetch_start_time) == int(watch.newest_history_key):
                                     logger.warning(
-                                        f"Timestamp {timestamp} already exists, waiting 1 seconds so we have a unique key in history.txt")
-                                    timestamp = str(int(timestamp) + 1)
+                                        f"Timestamp {fetch_start_time} already exists, waiting 1 seconds so we have a unique key in history.txt")
+                                    fetch_start_time += 1
                                     time.sleep(1)
 
                                 watch.save_history_text(contents=contents,
-                                                        timestamp=timestamp,
+                                                        timestamp=int(fetch_start_time),
                                                         snapshot_id=update_obj.get('previous_md5', 'none'))
 
-                                if update_handler.fetcher.content:
-                                    watch.save_last_fetched_html(contents=update_handler.fetcher.content, timestamp=timestamp)
 
-                                # Notifications should only trigger on the second time (first time, we gather the initial snapshot)
-                                if watch.history_n >= 2:
-                                    logger.info(f"Change detected in UUID {uuid} - {watch['url']}")
-                                    if not watch.get('notification_muted'):
-                                        self.send_content_changed_notification(watch_uuid=uuid)
+                                empty_pages_are_a_change = self.datastore.data['settings']['application'].get('empty_pages_are_a_change', False)
+                                if update_handler.fetcher.content or (not update_handler.fetcher.content and empty_pages_are_a_change):
+                                    # attribute .last_changed is then based on this data
+                                    watch.save_last_fetched_html(contents=update_handler.fetcher.content, timestamp=int(fetch_start_time))
+
+                                    # Notifications should only trigger on the second time (first time, we gather the initial snapshot)
+                                    if watch.history_n >= 2:
+                                        logger.info(f"Change detected in UUID {uuid} - {watch['url']}")
+                                        if not watch.get('notification_muted'):
+                                            self.send_content_changed_notification(watch_uuid=uuid)
 
                         except Exception as e:
                             # Catch everything possible here, so that if a worker crashes, we don't lose it until restart!
@@ -581,15 +585,15 @@ class update_worker(threading.Thread):
                     except Exception as e:
                         pass
 
-                    self.datastore.update_watch(uuid=uuid, update_obj={'fetch_time': round(time.time() - now, 3),
-                                                                       'last_checked': round(time.time()),
+                    self.datastore.update_watch(uuid=uuid, update_obj={'fetch_time': round(time.time() - fetch_start_time, 3),
+                                                                       'last_checked': int(fetch_start_time),
                                                                        'check_count': count
                                                                        })
 
 
                 self.current_uuid = None  # Done
                 self.q.task_done()
-                logger.debug(f"Watch {uuid} done in {time.time()-now:.2f}s")
+                logger.debug(f"Watch {uuid} done in {time.time()-fetch_start_time:.2f}s")
 
                 # Give the CPU time to interrupt
                 time.sleep(0.1)

docker-compose.yml

@@ -12,9 +12,6 @@ services:
   #    environment:
   #        Default listening port, can also be changed with the -p option
   #      - PORT=5000
-
-  #      - PUID=1000
-  #      - PGID=1000
   #
   #        Log levels are in descending order. (TRACE is the most detailed one)
   #        Log output levels: TRACE, DEBUG(default), INFO, SUCCESS, WARNING, ERROR, CRITICAL

requirements.txt

@@ -35,7 +35,7 @@ dnspython==2.6.1 # related to eventlet fixes
 # jq not available on Windows so must be installed manually
 
 # Notification library
-apprise==1.9.0
+apprise==1.9.1
 
 # apprise mqtt https://github.com/dgtlmoon/changedetection.io/issues/315
 # use any version other than 2.0.x due to https://github.com/eclipse/paho.mqtt.python/issues/814