feat(monitor_dir): Add path/inode polling mode option

monitor_dir.py

@@ -1,6 +1,6 @@
-#!/usr/bin/env -S uv run
+#!/usr/bin/env -S uv run --script
 # /// script
-# requires-python = ">=3.8"
+# requires-python = ">=3.11"
 # dependencies = [
 #   "watchdog"
 # ]
@@ -12,20 +12,20 @@ import logging
 import os
 from pathlib import Path
 import re
+import stat
 import subprocess
 import sys
 import time
 from typing import Self
-from watchdog import observers
 from watchdog.observers.polling import PollingObserver
 from watchdog import events
 
-logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s: %(message)s", datefmt="%Y-%m-%d %H:%M:%S")
+logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s: %(message)s")
 
 
 @dataclasses.dataclass
 class EventHandler:
-    pattern: re.Pattern
+    pattern: re.Pattern[str]
     script_path: Path
 
     @classmethod
@@ -51,24 +51,24 @@ class EventHandler:
 
 
 class ChangeDispatchHandler(events.FileSystemEventHandler):
-    def __init__(self, base_dir: Path, handlers_config: dict[str, EventHandler], ignore_patterns: list[re.Pattern]):
+    def __init__(self, base_dir: Path, handlers_config: dict[str, list[EventHandler]], ignore_patterns: list[re.Pattern[str]]) -> None:
         super().__init__()
         self.base_dir = base_dir
         self.handlers_config = handlers_config
         self.ignore_patterns = ignore_patterns if ignore_patterns else []
 
-    def _execute_script(self, script_path: str, event_name: str, src_path: str, dest_path: str = None):
+    def _execute_script(self, script_path: Path, event_name: str, src_path: Path, dest_path: Path | None = None) -> None:
         env_vars = os.environ.copy()
         env_vars["MONITORED_DIR"] = str(self.base_dir)
         env_vars["EVENT_TYPE"] = event_name.upper()
-        env_vars["SRC_PATH"] = str(Path(src_path).resolve())
+        env_vars["SRC_PATH"] = str(src_path.absolute())
         args = [
             event_name,
-            str(Path(src_path).resolve()),
-            *([str(Path(dest_path).resolve())] if dest_path else []),
+            str(src_path.absolute()),
+            *([str(dest_path.absolute())] if dest_path else []),
         ]
         if dest_path:
-            env_vars["DEST_PATH"] = str(Path(dest_path).resolve())
+            env_vars["DEST_PATH"] = str(dest_path.absolute())
         elif "DEST_PATH" in env_vars:
             del env_vars["DEST_PATH"]
 
@@ -93,7 +93,7 @@ class ChangeDispatchHandler(events.FileSystemEventHandler):
             logging.error(f"An unexpected error occurred while running script '{script_path}': {e}")
 
     def _is_ignored(self, path_to_check: Path) -> bool:
-        abs_path_to_check = path_to_check.resolve()
+        abs_path_to_check = path_to_check.absolute()
         for pattern in self.ignore_patterns:
             if pattern.search(str(abs_path_to_check)):
                 return True
@@ -107,7 +107,7 @@ class ChangeDispatchHandler(events.FileSystemEventHandler):
                 pass
         return False
 
-    def _process_event_type(self, event_name: str, path_to_filter_on: Path, src_path: Path, dest_path: Path | None = None):
+    def _process_event_type(self, event_name: str, path_to_filter_on: Path, src_path: Path, dest_path: Path | None = None) -> None:
         if self._is_ignored(path_to_filter_on):
             logging.debug(f"Path '{path_to_filter_on}' is ignored by configured patterns.")
             return
@@ -127,27 +127,27 @@ class ChangeDispatchHandler(events.FileSystemEventHandler):
                 dest_path=dest_path,
             )
 
-    def on_created(self, event: events.FileSystemEvent):
+    def on_created(self, event: events.FileSystemEvent) -> None:
         what = "directory" if event.is_directory else "file"
         logging.debug(f"Watchdog event: CREATED {what} at {event.src_path}")
 
         self._process_event_type(
-            event.event_type,
+            event_name=event.event_type,
             path_to_filter_on=Path(event.src_path),
             src_path=Path(event.src_path),
         )
 
-    def on_deleted(self, event: events.FileSystemEvent):
+    def on_deleted(self, event: events.FileSystemEvent) -> None:
         what = "directory" if event.is_directory else "file"
         logging.debug(f"Watchdog event: DELETED {what} at {event.src_path}")
 
         self._process_event_type(
-            event.event_type,
+            event_name=event.event_type,
             path_to_filter_on=Path(event.src_path),
             src_path=Path(event.src_path),
         )
 
-    def on_moved(self, event: events.FileSystemEvent):
+    def on_moved(self, event: events.FileSystemMovedEvent) -> None:
         what = "directory" if event.is_directory else "file"
         logging.debug(f"Watchdog event: MOVED {what} from {event.src_path} to {event.dest_path}")
 
@@ -155,19 +155,62 @@ class ChangeDispatchHandler(events.FileSystemEventHandler):
         dest_path = Path(event.dest_path)
 
         self._process_event_type(
-            event.event_type,
+            event_name=event.event_type,
             path_to_filter_on=dest_path,
             src_path=src_path,
             dest_path=dest_path,
         )
 
 
-def parse_args():
+def take_snapshot(base_dir: Path, recursive: bool) -> dict[Path, bool]:
+    if not stat.S_ISDIR(base_dir.stat().st_mode):
+        raise NotADirectoryError(f"Monitored path '{base_dir}' is not a directory")
+    snapshot: dict[Path, bool] = {}
+    directories = [base_dir]
+    while directories:
+        directory = directories.pop()
+        for path in directory.iterdir():
+            is_directory = stat.S_ISDIR(path.lstat().st_mode)
+            snapshot[path] = is_directory
+            if recursive and is_directory:
+                directories.append(path)
+    return snapshot
+
+
+def dispatch_changes(previous: dict[Path, bool], current: dict[Path, bool], event_handler: ChangeDispatchHandler) -> None:
+    deleted = previous.keys() - current.keys()
+    created = current.keys() - previous.keys()
+    replaced = {path for path in previous.keys() & current.keys() if previous[path] != current[path]}
+    for path in sorted(deleted | replaced):
+        event_handler._process_event_type(event_name="deleted", path_to_filter_on=path, src_path=path)
+    for path in sorted(created | replaced):
+        event_handler._process_event_type(event_name="created", path_to_filter_on=path, src_path=path)
+
+
+def monitor_paths(base_dir: Path, recursive: bool, interval: float, event_handler: ChangeDispatchHandler) -> None:
+    previous: dict[Path, bool] | None = None
+    while True:
+        try:
+            current = take_snapshot(base_dir=base_dir, recursive=recursive)
+        except OSError as error:
+            logging.warning(f"Directory scan failed; retaining previous snapshot and retrying in {interval}s: {error}")
+        else:
+            if previous is not None:
+                dispatch_changes(previous=previous, current=current, event_handler=event_handler)
+            previous = current
+        time.sleep(interval)
+
+
+def parse_args() -> argparse.Namespace:
     parser = argparse.ArgumentParser(
-        description="Monitors a folder for changes and executes handlers based on event type and regex filters.", formatter_class=argparse.RawTextHelpFormatter
+        description="Monitors a folder for changes and executes handlers based on event type and regex filters.", formatter_class=argparse.ArgumentDefaultsHelpFormatter
     )
     parser.add_argument("dir", type=Path, help="The directory to monitor.")
-    parser.add_argument("--poll", type=int, default=10, help="Polling interval in seconds.")
+    parser.add_argument("--poll", type=float, default=10, help="Polling interval in seconds; must be positive.")
+    parser.add_argument(
+        "--polling-mode", choices=["path", "inode"], default="path",
+        help="Path polling tolerates changing SSHFS inode numbers and retries failed scans. Renames emit deleted/created events. Inode polling supports moved events and requires stable inode numbers.",
+    )
     parser.add_argument("--recursive", action="store_true", help="Monitor the directory recursively.")
     parser.add_argument(
         "--created",
@@ -206,10 +249,15 @@ def parse_args():
         default=[],
     )
     parser.add_argument("--debug", action="store_true", help="Enable debug logging.")
-    return parser.parse_args()
+    args = parser.parse_args()
+    if not 0 < args.poll < float("inf"):
+        parser.error("--poll must be a finite positive number")
+    if args.moved_handlers and args.polling_mode != "inode":
+        parser.error("--moved requires --polling-mode inode; for SSHFS, use --created and --deleted instead")
+    return args
 
 
-def main():
+def main() -> None:
     args = parse_args()
 
     if args.debug:
@@ -220,7 +268,7 @@ def main():
         logging.error(f"Monitored directory '{args.dir}' (resolved to '{args.dir.resolve()}') not found or not a directory.")
         sys.exit(1)
 
-    handlers: dict[str, EventHandler] = {
+    handlers: dict[str, list[EventHandler]] = {
         "created": args.created_handlers,
         "deleted": args.deleted_handlers,
         "moved": args.moved_handlers,
@@ -235,7 +283,6 @@ def main():
         if handlers_list:
             logging.info(f"Registered {len(handlers_list)} handler(s) for '{event_type}' events.")
             for handler in handlers_list:
-                handler: EventHandler
                 logging.info(f"  - Filter: '{handler.pattern.pattern}', Script: '{handler.script_path}'")
 
     if args.ignore_patterns:
@@ -243,13 +290,22 @@ def main():
         for pattern in args.ignore_patterns:
             logging.info(f"  - {pattern.pattern}")
 
+    base_dir = args.dir.resolve()
     event_handler = ChangeDispatchHandler(
-        base_dir=args.dir.resolve(),
+        base_dir=base_dir,
         handlers_config=handlers,
         ignore_patterns=args.ignore_patterns,
     )
+    if args.polling_mode == "path":
+        logging.info("Monitor is now running with path polling. Press Ctrl+C to stop.")
+        try:
+            monitor_paths(base_dir=base_dir, recursive=args.recursive, interval=args.poll, event_handler=event_handler)
+        except KeyboardInterrupt:
+            logging.info("Monitor has been shut down.")
+        return
+
     observer = PollingObserver(timeout=args.poll)
-    observer.schedule(event_handler, args.dir, recursive=args.recursive)
+    observer.schedule(event_handler=event_handler, path=str(base_dir), recursive=args.recursive)
     observer.start()
     logging.info("Monitor is now running. Press Ctrl+C to stop.")
 

monitor_dir_test.py

@@ -0,0 +1,91 @@
+#!/usr/bin/env -S uv run --script
+# /// script
+# dependencies = ["pytest", "watchdog"]
+# ///
+
+from pathlib import Path
+import stat
+import sys
+from types import SimpleNamespace
+from unittest.mock import Mock, call, patch
+
+import pytest
+
+import monitor_dir as script
+
+
+@pytest.mark.parametrize("recursive", [False, True])
+def test_recursive_scan_does_not_follow_symlinks(tmp_path: Path, recursive: bool) -> None:
+    child = tmp_path / "child"
+    child.mkdir()
+    file_path = child / "file.txt"
+    file_path.touch()
+    link = tmp_path / "loop"
+    link.symlink_to(target=tmp_path, target_is_directory=True)
+
+    expected = {child: True, link: False}
+    if recursive:
+        expected[file_path] = False
+    assert script.take_snapshot(base_dir=tmp_path, recursive=recursive) == expected
+
+
+def test_changing_inode_numbers_do_not_emit_events(tmp_path: Path) -> None:
+    file_path = tmp_path / "file.txt"
+    file_path.touch()
+    handler = Mock()
+    with patch.object(
+        target=Path, attribute="lstat",
+        side_effect=[
+            SimpleNamespace(st_mode=stat.S_IFREG, st_ino=1),
+            SimpleNamespace(st_mode=stat.S_IFREG, st_ino=999),
+        ],
+    ):
+        previous = script.take_snapshot(base_dir=tmp_path, recursive=False)
+        current = script.take_snapshot(base_dir=tmp_path, recursive=False)
+
+    script.dispatch_changes(previous=previous, current=current, event_handler=handler)
+    handler._process_event_type.assert_not_called()
+
+
+def test_rename_emits_deleted_and_created() -> None:
+    source = Path("/watched/before.txt")
+    destination = Path("/watched/after.txt")
+    handler = Mock()
+    script.dispatch_changes(previous={source: False}, current={destination: False}, event_handler=handler)
+    assert handler._process_event_type.call_args_list == [
+        call(event_name="deleted", path_to_filter_on=source, src_path=source),
+        call(event_name="created", path_to_filter_on=destination, src_path=destination),
+    ]
+
+
+def test_failed_scan_preserves_last_snapshot() -> None:
+    base_dir = Path("/watched")
+    existing = base_dir / "existing.txt"
+    created = base_dir / "created.txt"
+    handler = Mock()
+    with patch.object(
+        target=script, attribute="take_snapshot",
+        side_effect=[{existing: False}, OSError("SSHFS disconnected"), {existing: False, created: False}],
+    ), patch.object(
+        target=script.time, attribute="sleep", side_effect=[None, None, KeyboardInterrupt],
+    ):
+        with pytest.raises(KeyboardInterrupt):
+            script.monitor_paths(base_dir=base_dir, recursive=True, interval=1, event_handler=handler)
+    handler._process_event_type.assert_called_once_with(
+        event_name="created", path_to_filter_on=created, src_path=created,
+    )
+
+
+def test_partial_scan_is_discarded(tmp_path: Path) -> None:
+    (tmp_path / "first.txt").touch()
+    (tmp_path / "second.txt").touch()
+    with patch.object(
+        target=Path, attribute="lstat",
+        side_effect=[SimpleNamespace(st_mode=stat.S_IFREG), PermissionError("scan interrupted")],
+    ):
+        with pytest.raises(PermissionError):
+            script.take_snapshot(base_dir=tmp_path, recursive=True)
+
+
+if __name__ == "__main__":
+    sys.exit(pytest.main(args=[str(Path(__file__).resolve()), *sys.argv[1:]]))