change README.md

README.md

@@ -1,5 +1,5 @@
 # pyinotifyd
-A daemon to monitor filesystems events with inotify on Linux and run tasks like filesystem operations (copy, move or delete), a shell command or custom async python methods.  
+A daemon for monitoring filesystem events with inotify on Linux and run tasks like filesystem operations (copy, move or delete), a shell command or custom async python methods.  
 
 It is possible to schedule tasks with a delay, which can then be canceled again in case a canceling event occurs. A useful example for this is to run tasks only if a file has not changed within a certain amount of time.  
 
@@ -49,7 +49,7 @@ The basic idea is to instantiate one or multiple schedulers and map specific ino
 pyinotifyd has different schedulers to schedule tasks with an optional delay. The advantages of using a scheduler are consistent logging and the possibility to cancel delayed tasks. Furthermore, schedulers have the ability to differentiate between files and directories.
 
 ### TaskScheduler
-Schedule a custom python method *job* with an optional *delay* in seconds. Skip scheduling of tasks for files and/or directories according to *files* and *dirs* arguments. If there already is a scheduled task, re-schedule it with *delay*. Use *logname* in log messages. All additional modules, functions and variables that are defined in the config file and are needed within the *job*, need to be passed as dictionary to the TaskManager through *global_vars*.  
+Schedule a custom python method *job* with an optional *delay* in seconds. Skip scheduling of tasks for files and/or directories according to *files* and *dirs* arguments. If there already is a scheduled task, re-schedule it with *delay*. Use *logname* in log messages. All additional modules, functions and variables that are defined in the config file and are needed within the *job*, need to be passed as dictionary to the TaskManager through *global_vars*. If you want to limit the scheduler to run only one job at a time, set *singlejob* to True.  
 All arguments except for *job* are optional.
 ```python
 # Please note that pyinotifyd uses pythons asyncio for asynchronous task execution.
@@ -71,7 +71,8 @@ task_sched = TaskScheduler(
     dirs=False,
     delay=0,
     logname="sched",
-    global_vars=globals())
+    global_vars=globals(),
+    singlejob=False)
 ```
 
 ### ShellScheduler
@@ -94,8 +95,8 @@ If *action* is delete, delete non-empty directories if *rec* is set to True.
 ```python
 move_rule = FileManagerRule(
     action="move",
-    src_re="^/src_path/(?P<path>.*).to_move$",
+    src_re=r"^/src_path/(?P<path>.*).to_move$",
-    dst_re="/dst_path/\g<path>",
+    dst_re=r"/dst_path/\g<path>",
     auto_create=False,
     rec=False,
     filemode=None,
@@ -106,7 +107,7 @@ move_rule = FileManagerRule(
 
 delete_rule = FileManagerRule(
     action="delete",
-    src_re="^/src_path/(?P<path>.*).to_delete$",
+    src_re=r"^/src_path/(?P<path>.*).to_delete$",
     rec=False)
 
 file_sched = FileManagerScheduler(
@@ -164,21 +165,23 @@ pyinotifyd = Pyinotifyd(
 ```
 
 ### Watches
-A watch connects the *path* to an *event_map*. Automatically add a watch on each sub-directories in *path* if *rec* is set to True. If *auto_add* is True, a watch will be added automatically on newly created sub-directories in *path*.
+A watch connects the *path* to an *event_map*. Automatically add a watch on each sub-directories in *path* if *rec* is set to True. If *auto_add* is True, a watch will be added automatically on newly created sub-directories in *path*. All events for paths matching one of the regular expressions in *exclude_filter* are ignored. If the value of *exclude_filter* is a string, it is assumed to be a path to a file from which the list of regular expressions will be read.
 ```python
 # Add a watch directly to Pyinotifyd.
 pyinotifyd.add_watch(
     path="/src_path",
     event_map=event_map,
     rec=False,
-    auto_add=False)
+    auto_add=False,
+    exclude_filter=["^/src_path/subpath$"])
 
 # Or instantiate and add it
 w = Watch(
     path="/src_path",
     event_map=event_map,
     rec=False,
-    auto_add=False)
+    auto_add=False,
+    exclude_filter=["^/src_path/subpath$"])
 
 pyinotifyd.add_watch(watch=w)
 ```
@@ -214,6 +217,8 @@ enableSyslog(lglevel=INFO, name="daemon")
 
 ## Schedule python method for all events on files and directories
 ```python
+import logging
+
 async def custom_job(event, task_id):
     logging.info(f"{task_id}: execute example task: {event}")
 
@@ -255,23 +260,23 @@ pyinotifyd.add_watch(
 ```python
 move_rule = FileManagerRule(
     action="move",
-    src_re="^/src_path/(?P<path>.*)\.to_move$",
+    src_re=r"^/src_path/(?P<path>.*)\.to_move$",
-    dst_re="/dst_path/\g<path>",
+    dst_re=r"/dst_path/\g<path>",
     auto_create=True,
     filemode=0o644,
     dirmode=0o755)
 
 copy_rule = FileManagerRule(
     action="copy",
-    src_re="^/src_path/(?P<path>.*)\.to_copy$",
+    src_re=r"^/src_path/(?P<path>.*)\.to_copy$",
-    dst_re="/dst_path/\g<path>",
+    dst_re=r"/dst_path/\g<path>",
     auto_create=True,
     filemode=0o644,
     dirmode=0o755)
 
 delete_rule = FileManagerRule(
     action="delete",
-    src_re="^/src_path/(?P<path>.*)\.to_delete$",
+    src_re=r"^/src_path/(?P<path>.*)\.to_delete$",
     rec=False)
 
 file_sched = FileManagerScheduler(

distribution/gentoo/dev-python/pyinotifyd/pyinotifyd-0.0.6.ebuild

@@ -1,47 +0,0 @@
-# Copyright 2020 Gentoo Authors
-# Distributed under the terms of the GNU General Public License v2
-
-EAPI=7
-PYTHON_COMPAT=( python3_{8..10} )
-DISTUTILS_USE_SETUPTOOLS=rdepend
-
-SCM=""
-if [ "${PV#9999}" != "${PV}" ] ; then
-	SCM="git-r3"
-	EGIT_REPO_URI="https://github.com/spacefreak86/${PN}"
-fi
-
-inherit ${SCM} distutils-r1 systemd
-
-DESCRIPTION="Monitore filesystems events and execute Python methods or Shell commands."
-HOMEPAGE="https://github.com/spacefreak86/pymodmilter"
-if [ "${PV#9999}" != "${PV}" ] ; then
-	SRC_URI=""
-	KEYWORDS=""
-	# Needed for tests
-	S="${WORKDIR}/${PN}"
-	EGIT_CHECKOUT_DIR="${S}"
-else
-	SRC_URI="https://github.com/spacefreak86/${PN}/archive/${PV}.tar.gz -> ${P}.tar.gz"
-	KEYWORDS="amd64 x86"
-fi
-
-LICENSE="GPL-3"
-SLOT="0"
-
-IUSE="systemd"
-
-RDEPEND="dev-python/pyinotify[${PYTHON_USEDEP}]"
-
-python_install_all() {
-	distutils-r1_python_install_all
-
-	dodir /etc/${PN}
-	insinto /etc/${PN}
-	newins ${PN}/misc/config.py.default config.py
-
-	use systemd && systemd_dounit ${PN}/misc/${PN}.service
-
-	newinitd ${PN}/misc/openrc/${PN}.initd ${PN}
-	newconfd ${PN}/misc/openrc/${PN}.confd ${PN}
-}

distribution/gentoo/dev-python/pyinotifyd/pyinotifyd-9999.ebuild

@@ -1,9 +1,9 @@
 # Copyright 2020 Gentoo Authors
 # Distributed under the terms of the GNU General Public License v2
 
-EAPI=7
+EAPI=8
-PYTHON_COMPAT=( python3_{7,8,9} )
+PYTHON_COMPAT=( python3_{13..14} )
-DISTUTILS_USE_SETUPTOOLS=rdepend
+DISTUTILS_USE_PEP517=setuptools
 
 SCM=""
 if [ "${PV#9999}" != "${PV}" ] ; then

pyinotifyd/__init__.py

@@ -31,12 +31,12 @@ import pyinotify
 import signal
 import sys
 
-from pyinotify import ProcessEvent
+from pyinotify import ProcessEvent, ExcludeFilter
 
 from pyinotifyd._install import install, uninstall
 from pyinotifyd.scheduler import TaskScheduler, Cancel
 
-__version__ = "0.0.6"
+__version__ = "1.0.2"
 
 
 def setLoglevel(loglevel, logname=None):
@@ -56,16 +56,15 @@ def enableSyslog(loglevel=None, address="/dev/log", logname=None):
 
 
 class _SchedulerList:
-    def __init__(self, schedulers=[], loop=None):
+    def __init__(self, schedulers=[]):
         if not isinstance(schedulers, list):
             schedulers = [schedulers]
 
         self._schedulers = schedulers
-        self._loop = (loop or asyncio.get_event_loop())
 
     def process_event(self, event):
         for scheduler in self._schedulers:
-            self._loop.create_task(scheduler.process_event(event))
+            asyncio.create_task(scheduler.process_event(event))
 
     def schedulers(self):
         return self._schedulers
@@ -76,10 +75,10 @@ class EventMap(ProcessEvent):
         **pyinotify.EventsCodes.OP_FLAGS,
         **pyinotify.EventsCodes.EVENT_FLAGS}
 
-    def my_init(self, event_map=None, default_sched=None, loop=None,
+    def my_init(self, event_map=None, default_sched=None, exclude_filter=None,
                 logname="eventmap"):
         self._map = {}
-        self._loop = (loop or asyncio.get_event_loop())
+        self._exclude_filter = None
 
         if default_sched is not None:
             for flag in EventMap.flags:
@@ -91,6 +90,7 @@ class EventMap(ProcessEvent):
             for flag, schedulers in event_map.items():
                 self.set_scheduler(flag, schedulers)
 
+        self.set_exclude_filter(exclude_filter)
         self._log = logging.getLogger((logname or __name__))
 
     def set_scheduler(self, flag, schedulers):
@@ -106,28 +106,44 @@ class EventMap(ProcessEvent):
                         isinstance(scheduler, Cancel):
                     instances.append(scheduler)
                 else:
-                    instances.append(
+                    instances.append(TaskScheduler(scheduler))
-                        TaskScheduler(scheduler, loop=self._loop))
 
-            self._map[flag] = _SchedulerList(instances, loop=self._loop)
+            self._map[flag] = _SchedulerList(instances)
 
         elif flag in self._map:
             del self._map[flag]
 
+    def set_exclude_filter(self, exclude_filter):
+        if exclude_filter is None:
+            self._exclude_filter = None
+            return
+
+        if not isinstance(exclude_filter, ExcludeFilter):
+            self._exclude_filter = ExcludeFilter(exclude_filter)
+        else:
+            self._exclude_filter = exclude_filter
+
     def process_default(self, event):
-        msg = "received event"
+        attrs = ""
         for attr in [
                 "dir", "mask", "maskname", "pathname", "src_pathname", "wd"]:
             value = getattr(event, attr, None)
             if attr == "mask":
                 value = hex(value)
             if value:
-                msg += f", {attr}={value}"
+                attrs += f", {attr}={value}"
 
-        self._log.debug(msg)
+        self._log.debug(f"received event{attrs}")
         maskname = event.maskname.split("|")[0]
-        if maskname in self._map:
+
-            self._map[maskname].process_event(event)
+        if maskname not in self._map:
+            return
+
+        if self._exclude_filter and self._exclude_filter(event.pathname):
+            self._log.debug(f"pathname {event.pathname} is excluded")
+            return
+
+        self._map[maskname].process_event(event)
 
     def schedulers(self):
         schedulers = []
@@ -139,23 +155,32 @@ class EventMap(ProcessEvent):
 
 
 class Watch:
-    def __init__(self, path, event_map=None, default_sched=None, rec=False,
+    def __init__(self, path, event_map=None, default_sched=None,
-                 auto_add=False, logname="watch", loop=None):
+                 rec=False, auto_add=False, exclude_filter=None,
-        assert isinstance(path, str), \
+                 logname="watch"):
-            f"path: expected {type('')}, got {type(path)}"
+        assert (isinstance(path, str) or isinstance(path, list)), \
+            f"path: expected {type('')} or {type([])}, got {type(path)}"
 
         if isinstance(event_map, EventMap):
             self._event_map = event_map
         else:
             self._event_map = EventMap(
-                event_map=event_map, default_sched=default_sched)
+                event_map=event_map, default_sched=default_sched,
+                exclude_filter=exclude_filter)
 
         assert isinstance(rec, bool), \
             f"rec: expected {type(bool)}, got {type(rec)}"
         assert isinstance(auto_add, bool), \
             f"auto_add: expected {type(bool)}, got {type(auto_add)}"
+
+        self._exclude_filter = None
+        if exclude_filter:
+            if not isinstance(exclude_filter, ExcludeFilter):
+                self._exclude_filter = ExcludeFilter(exclude_filter)
+            else:
+                self._exclude_filter = exclude_filter
+
         logname = (logname or __name__)
-        self._loop = loop
 
         self._path = path
         self._rec = rec
@@ -171,14 +196,16 @@ class Watch:
     def event_map(self):
         return self._event_map
 
-    def start(self, loop=None):
+    def start(self):
-        loop = (loop or self._loop)
         self._watch_manager.add_watch(self._path, pyinotify.ALL_EVENTS,
                                       rec=self._rec, auto_add=self._auto_add,
+                                      exclude_filter=self._exclude_filter,
                                       do_glob=True)
 
         self._notifier = pyinotify.AsyncioNotifier(
-            self._watch_manager, loop, default_proc_fun=self._event_map)
+                self._watch_manager,
+                asyncio.get_event_loop(),
+                default_proc_fun=self._event_map)
 
     def stop(self):
         self._notifier.stop()
@@ -189,12 +216,10 @@ class Watch:
 class Pyinotifyd:
     name = "pyinotifyd"
 
-    def __init__(self, watches=[], shutdown_timeout=30, logname="daemon",
+    def __init__(self, watches=[], shutdown_timeout=30, logname="daemon"):
-                 loop=None):
         self.set_watches(watches)
         self.set_shutdown_timeout(shutdown_timeout)
         logname = (logname or __name__)
-        self._loop = (loop or asyncio.get_event_loop())
 
         self._log = logging.getLogger(logname)
 
@@ -203,7 +228,7 @@ class Pyinotifyd:
         config = {}
         name = Pyinotifyd.name
         exec("from logging import DEBUG, INFO, WARNING, ERROR, CRITICAL",
-            config)
+             config)
         exec(f"from {name} import Pyinotifyd, Watch", config)
         exec(f"from {name} import setLoglevel, enableSyslog", config)
         exec(f"from {name}.scheduler import *", config)
@@ -246,9 +271,7 @@ class Pyinotifyd:
             schedulers.extend(w.event_map().schedulers())
         return list(set(schedulers))
 
-    def start(self, loop=None):
+    def start(self):
-        loop = (loop or self._loop)
-
         if len(self._watches) == 0:
             self._log.warning(
                 "no watches configured, the daemon will not do anything")
@@ -256,7 +279,7 @@ class Pyinotifyd:
         for watch in self._watches:
             self._log.info(
                 f"start listening for inotify events on '{watch.path()}'")
-            watch.start(loop)
+            watch.start()
 
     def pause(self):
         for scheduler in self.schedulers():
@@ -435,7 +458,7 @@ def main():
         f"%(asctime)s - {name}/%(name)s - %(levelname)s - %(message)s")
     ch.setFormatter(formatter)
 
-    loop = asyncio.get_event_loop()
+    loop = asyncio.new_event_loop()
     loop.add_signal_handler(
         signal.SIGTERM, lambda: loop.create_task(
             daemon.shutdown("SIGTERM")))
@@ -445,6 +468,7 @@ def main():
     loop.add_signal_handler(
         signal.SIGHUP, lambda: loop.create_task(
             daemon.reload("SIGHUP", args.config, args.debug)))
+    asyncio.set_event_loop(loop)
 
     daemon.start()
     loop.run_forever()

pyinotifyd/misc/config.py.default

@@ -6,15 +6,16 @@
 #import logging
 #
 #async def custom_job(event, task_id):
-#    asyncio.sleep(1)
+#    await asyncio.sleep(1)
 #    logging.info(f"{task_id}: execute example task: {event}")
 #
 #task_sched = TaskScheduler(
 #    job=custom_job,
 #    files=True,
 #    dirs=False,
-#    delay=10
+#    delay=10,
-#    global_vars=globals())
+#    global_vars=globals(),
+#    singlejob=False)
 
 
 ###########################
@@ -25,7 +26,8 @@
 #    cmd="/usr/local/bin/task.sh {maskname} {pathname} {src_pathname}",
 #    files=True,
 #    dirs=False,
-#    delay=10)
+#    delay=10,
+#    singlejob=False)
 
 
 #################################
@@ -34,8 +36,8 @@
 
 #move_rule = Rule(
 #    action="move",
-#    src_re="^/src_path/(?P<path>.*).to_move",
+#    src_re=r"^/src_path/(?P<path>.*).to_move",
-#    dst_re="/dst_path/\g<path>.moved",
+#    dst_re=r"/dst_path/\g<path>.moved",
 #    auto_create=True,
 #    filemode=0o755,
 #    dirmode=0o644,
@@ -45,7 +47,7 @@
 
 #delete_rule = Rule(
 #    action="delete",
-#    src_re="^/src_path/(?P<path>.*).to_delete",
+#    src_re=r"^/src_path/(?P<path>.*).to_delete",
 #    rec=False)
 
 #file_sched = FileManagerScheduler(
@@ -86,7 +88,8 @@
 #    path="/watched/directory",
 #    event_map = event_map,
 #    rec=True,
-#    auto_add=True)
+#    auto_add=True,
+#    exclude_filter=["^/watched/directory/subpath$"])
 
 
 ################

pyinotifyd/scheduler.py

@@ -52,7 +52,7 @@ class TaskScheduler:
             self.cancelable = cancelable
 
     def __init__(self, job, files=True, dirs=False, delay=0, logname="sched",
-                 loop=None, global_vars={}):
+                 global_vars={}, singlejob=False):
         assert iscoroutinefunction(job), \
             f"job: expected coroutine, got {type(job)}"
         assert isinstance(files, bool), \
@@ -69,8 +69,8 @@ class TaskScheduler:
         self._dirs = dirs
         self._delay = delay
         self._log = logging.getLogger((logname or __name__))
-        self._loop = (loop or asyncio.get_event_loop())
         self._globals = global_vars
+        self._singlejob = singlejob
         self._tasks = {}
         self._pause = False
 
@@ -90,8 +90,7 @@ class TaskScheduler:
                 self._log.info(
                     f"wait {timeout} seconds for {len(pending)} "
                     f"remaining task(s) to complete")
-            done, pending = await asyncio.wait([*pending], timeout=timeout,
+            done, pending = await asyncio.wait([*pending], timeout=timeout)
-                                               loop=self._loop)
             if pending:
                 self._log.warning(
                     f"shutdown timeout exceeded, "
@@ -99,20 +98,23 @@ class TaskScheduler:
                 for task in pending:
                     task.cancel()
                 try:
-                    await asyncio.gather(*pending, loop=self._loop)
+                    await asyncio.gather(*pending)
                 except asyncio.CancelledError:
                     pass
             else:
                 self._log.info("all remainig tasks completed")
 
+    def taskindex(self, event):
+        return "singlejob" if self._singlejob else event.pathname
+
     async def _run_job(self, event, task_state, restart=False):
         logger = SchedulerLogger(self._log, {
             "event": event,
             "id": task_state.id})
 
         if self._delay > 0:
-            task_state.task = self._loop.create_task(
+            task_state.task = asyncio.create_task(
-                asyncio.sleep(self._delay, loop=self._loop))
+                asyncio.sleep(self._delay))
             try:
                 if restart:
                     prefix = "re-"
@@ -130,11 +132,11 @@ class TaskScheduler:
             local_vars = {"self": self,
                           "event": event,
                           "task_id": task_state.id}
-            task_state.task = self._loop.create_task(
+            task_state.task = asyncio.create_task(
                 eval("self._job(event, task_id)", self._globals, local_vars))
 
         else:
-            task_state.task = self._loop.create_task(
+            task_state.task = asyncio.create_task(
                 self._job(event, task_state.id))
 
         try:
@@ -145,7 +147,8 @@ class TaskScheduler:
         else:
             logger.info("task finished")
         finally:
-            del self._tasks[event.pathname]
+            task_index = self.taskindex(event)
+            del self._tasks[task_index]
 
     async def process_event(self, event):
         if not ((not event.dir and self._files) or
@@ -153,9 +156,13 @@ class TaskScheduler:
             return
 
         restart = False
+        task_index = self.taskindex(event)
         try:
-            task_state = self._tasks[event.pathname]
+            task_state = self._tasks[task_index]
-
+        except KeyError:
+            task_state = TaskScheduler.TaskState()
+            self._tasks[task_index] = task_state
+        else:
             logger = SchedulerLogger(self._log, {
                 "event": event,
                 "id": task_state.id})
@@ -171,16 +178,13 @@ class TaskScheduler:
                 logger.warning("skip event due to ongoing task")
                 return
 
-        except KeyError:
-            task_state = TaskScheduler.TaskState()
-            self._tasks[event.pathname] = task_state
-
         if not self._pause:
             await self._run_job(event, task_state, restart)
 
     async def process_cancel_event(self, event):
         try:
-            task_state = self._tasks[event.pathname]
+            task_index = self.taskindex(event)
+            task_state = self._tasks[task_index]
         except KeyError:
             return
 
@@ -192,7 +196,8 @@ class TaskScheduler:
             task_state.task.cancel()
             logger.info("scheduled task cancelled")
             task_state.task = None
-            del self._tasks[event.pathname]
+            logger.info(f"{task_index}")
+            del self._tasks[task_index]
         else:
             logger.warning("skip event due to ongoing task")
 
@@ -237,7 +242,7 @@ class ShellScheduler(TaskScheduler):
 
         logger.info(f"execute shell command, cmd={cmd}")
         try:
-            proc = await asyncio.create_subprocess_shell(cmd, loop=self._loop)
+            proc = await asyncio.create_subprocess_shell(cmd)
             await proc.communicate()
         except Exception as e:
             logger.error(e)
@@ -285,7 +290,8 @@ class FileManagerRule:
 
 class FileManagerScheduler(TaskScheduler):
     def __init__(self, rules, job=None, *args, **kwargs):
-        super().__init__(*args, **kwargs, job=self._manager_job)
+        super().__init__(
+            *args, **kwargs, job=self._manager_job, singlejob=False)
 
         if not isinstance(rules, list):
             rules = [rules]