From 0a3fd4065a45aacd294c8a2d48895d5ac3c6b7a1 Mon Sep 17 00:00:00 2001
From: Fabio Manganiello <info@fabiomanganiello.com>
Date: Mon, 20 Dec 2021 00:58:41 +0100
Subject: [PATCH] Proper implementation for filesystem monitor filters

The logic Watchdog applies to filter events based
on `ignore_directories`, `ignore_patterns` and
`ignore_regexes` isn't really sophisticated, and
it doesn't check whether a partial directory/file
name is used in one of the `ignore_*` patterns.
The `file.monitor` backend should therefore implement
this logic on its side.
---
 .../backend/file/monitor/entities/handlers.py | 43 +++++++++++++++++--
 1 file changed, 40 insertions(+), 3 deletions(-)

diff --git a/platypush/backend/file/monitor/entities/handlers.py b/platypush/backend/file/monitor/entities/handlers.py
index 738bca56b..ef7930317 100644
--- a/platypush/backend/file/monitor/entities/handlers.py
+++ b/platypush/backend/file/monitor/entities/handlers.py
@@ -1,4 +1,5 @@
 import os
+import re
 
 from watchdog.events import FileSystemEventHandler, PatternMatchingEventHandler, RegexMatchingEventHandler
 
@@ -16,14 +17,50 @@ class EventHandler(FileSystemEventHandler):
         resource.path = os.path.expanduser(resource.path)
         self.resource = resource
 
+    def _should_ignore_event(self, event) -> bool:
+        ignore_dirs = [
+            os.path.expanduser(
+                _dir if os.path.expanduser(_dir).strip('/').startswith(self.resource.path)
+                else os.path.join(self.resource.path, _dir)
+            )
+            for _dir in getattr(self.resource, 'ignore_directories', [])
+        ]
+
+        ignore_patterns = getattr(self.resource, 'ignore_patterns', None)
+        ignore_regexes = getattr(self.resource, 'ignore_regexes', None)
+
+        if ignore_dirs and any(
+                event.src_path.startswith(ignore_dir) for ignore_dir in ignore_dirs
+        ):
+            return True
+
+        if ignore_patterns and any(
+                re.match(r'^{}$'.format(pattern.replace('*', '.*')), event.src_path)
+                for pattern in ignore_patterns
+        ):
+            return True
+
+        if ignore_regexes and any(
+                re.match(regex, event.src_path)
+                for regex in ignore_patterns
+        ):
+            return True
+
+        return False
+
+    def _on_event(self, event, output_event_type):
+        if self._should_ignore_event(event):
+            return
+        get_bus().post(output_event_type(path=event.src_path, is_directory=event.is_directory))
+
     def on_created(self, event):
-        get_bus().post(FileSystemCreateEvent(path=event.src_path, is_directory=event.is_directory))
+        self._on_event(event, FileSystemCreateEvent)
 
     def on_deleted(self, event):
-        get_bus().post(FileSystemDeleteEvent(path=event.src_path, is_directory=event.is_directory))
+        self._on_event(event, FileSystemDeleteEvent)
 
     def on_modified(self, event):
-        get_bus().post(FileSystemModifyEvent(path=event.src_path, is_directory=event.is_directory))
+        self._on_event(event, FileSystemModifyEvent)
 
     def on_moved(self, event):
         pass