[fix] prevent multiple, parallel initializations of tables in the cache DB (#4991)

Depending on the respective runtime behavior, it could happen that the initial
loading of the DB tables in the cache was performed multiple times and in
parallel. The concurrent accesses then led to the `sqlite3.OperationalError:
database is locked` exception as in #4951.

Since this problem depends significantly on the runtimes (e.g., how long it
takes to retrieve the content for a table), this error could not be observed in
all installations.

Closes: https://github.com/searxng/searxng/issues/4951

Signed-off-by: Markus Heiser <markus.heiser@darmarit.de>

searx/data/tracker_patterns.py

@@ -7,7 +7,6 @@ import typing
 __all__ = ["TrackerPatternsDB"]
 
 import re
-import pathlib
 from collections.abc import Iterator
 from urllib.parse import urlparse, urlunparse, parse_qsl, urlencode
 
@@ -22,7 +21,6 @@ class TrackerPatternsDB:
     # pylint: disable=missing-class-docstring
 
     ctx_name = "data_tracker_patterns"
-    json_file = pathlib.Path(__file__).parent / "tracker_patterns.json"
 
     CLEAR_LIST_URL = [
         # ClearURL rule lists, the first one that responds HTTP 200 is used
@@ -42,8 +40,9 @@ class TrackerPatternsDB:
 
     def init(self):
         if self.cache.properties("tracker_patterns loaded") != "OK":
-            self.load()
+            # To avoid parallel initializations, the property is set first
             self.cache.properties.set("tracker_patterns loaded", "OK")
+            self.load()
         # F I X M E:
         #     do we need a maintenance .. rember: database is stored
         #     in /tmp and will be rebuild during the reboot anyway