Merge pull request #167 from bobloy/fifo_develop

FIFO expired triggers

fifo/date_trigger.py

@@ -0,0 +1,7 @@
+from apscheduler.triggers.date import DateTrigger
+
+
+class CustomDateTrigger(DateTrigger):
+    def get_next_fire_time(self, previous_fire_time, now):
+        next_run = super().get_next_fire_time(previous_fire_time, now)
+        return next_run if next_run >= now else None

fifo/fifo.py

@@ -4,6 +4,7 @@ from datetime import MAXYEAR, datetime, timedelta, tzinfo
 from typing import Optional, Union
 
 import discord
+import pytz
 from apscheduler.job import Job
 from apscheduler.jobstores.base import JobLookupError
 from apscheduler.schedulers.asyncio import AsyncIOScheduler
@@ -51,7 +52,7 @@ def _get_run_times(job: Job, now: datetime = None):
 
     if now is None:
         now = datetime(MAXYEAR, 12, 31, 23, 59, 59, 999999, tzinfo=job.next_run_time.tzinfo)
-        yield from _get_run_times(job, now)
+        yield from _get_run_times(job, now)  # Recursion
         raise StopIteration()
 
     next_run_time = job.next_run_time
@@ -145,28 +146,38 @@ class FIFO(commands.Cog):
         await task.delete_self()
 
     async def _process_task(self, task: Task):
-        job: Union[Job, None] = await self._get_job(task)
+        # None of this is necessar, we have `replace_existing` already
-        if job is not None:
+        # job: Union[Job, None] = await self._get_job(task)
-            job.reschedule(await task.get_combined_trigger())
+        # if job is not None:
-            return job
+        #     combined_trigger_ = await task.get_combined_trigger()
+        #     if combined_trigger_ is None:
+        #         job.remove()
+        #     else:
+        #         job.reschedule(combined_trigger_)
+        #     return job
         return await self._add_job(task)
 
     async def _get_job(self, task: Task) -> Job:
         return self.scheduler.get_job(_assemble_job_id(task.name, task.guild_id))
 
     async def _add_job(self, task: Task):
+        combined_trigger_ = await task.get_combined_trigger()
+        if combined_trigger_ is None:
+            return None
+
         return self.scheduler.add_job(
             _execute_task,
             kwargs=task.__getstate__(),
             id=_assemble_job_id(task.name, task.guild_id),
-            trigger=await task.get_combined_trigger(),
+            trigger=combined_trigger_,
             name=task.name,
         )
 
     async def _resume_job(self, task: Task):
-        try:
+        job: Union[Job, None] = await self._get_job(task)
-            job = self.scheduler.resume_job(job_id=_assemble_job_id(task.name, task.guild_id))
+        if job is not None:
-        except JobLookupError:
+            job.resume()
+        else:
             job = await self._process_task(task)
         return job
 
@@ -221,6 +232,17 @@ class FIFO(commands.Cog):
         if ctx.invoked_subcommand is None:
             pass
 
+    @fifo.command(name="wakeup")
+    async def fifo_wakeup(self, ctx: commands.Context):
+        """Debug command to fix missed executions.
+
+        If you see a negative "Next run time" when adding a trigger, this may help resolve it.
+        Check the logs when using this command.
+        """
+
+        self.scheduler.wakeup()
+        await ctx.tick()
+
     @fifo.command(name="checktask", aliases=["checkjob", "check"])
     async def fifo_checktask(self, ctx: commands.Context, task_name: str):
         """Returns the next 10 scheduled executions of the task"""
@@ -372,10 +394,14 @@ class FIFO(commands.Cog):
 
         else:
             embed.add_field(name="Server", value="Server not found", inline=False)
+        triggers, expired_triggers = await task.get_triggers()
 
-        trigger_str = "\n".join(str(t) for t in await task.get_triggers())
+        trigger_str = "\n".join(str(t) for t in triggers)
+        expired_str = "\n".join(str(t) for t in expired_triggers)
         if trigger_str:
             embed.add_field(name="Triggers", value=trigger_str, inline=False)
+        if expired_str:
+            embed.add_field(name="Expired Triggers", value=expired_str, inline=False)
 
         job = await self._get_job(task)
         if job and job.next_run_time:
@@ -546,7 +572,7 @@ class FIFO(commands.Cog):
             )
             return
 
-        time_to_run = datetime.now() + time_from_now
+        time_to_run = datetime.now(pytz.utc) + time_from_now
 
         result = await task.add_trigger("date", time_to_run, time_to_run.tzinfo)
         if not result:

fifo/redconfigjobstore.py

@@ -39,7 +39,7 @@ class RedConfigJobStore(MemoryJobStore):
         # self._jobs = [
         #     (await self._decode_job(job), timestamp) async for (job, timestamp) in AsyncIter(_jobs)
         # ]
-        async for job, timestamp in AsyncIter(_jobs):
+        async for job, timestamp in AsyncIter(_jobs, steps=5):
             job = await self._decode_job(job)
             index = self._get_job_index(timestamp, job.id)
             self._jobs.insert(index, (job, timestamp))
@@ -109,83 +109,6 @@ class RedConfigJobStore(MemoryJobStore):
 
         return job
 
-    # @run_in_event_loop
-    # def add_job(self, job: Job):
-    #     if job.id in self._jobs_index:
-    #         raise ConflictingIdError(job.id)
-    #     # log.debug(f"Check job args: {job.args=}")
-    #     timestamp = datetime_to_utc_timestamp(job.next_run_time)
-    #     index = self._get_job_index(timestamp, job.id)  # This is fine
-    #     self._jobs.insert(index, (job, timestamp))
-    #     self._jobs_index[job.id] = (job, timestamp)
-    #     task = asyncio.create_task(self._async_add_job(job, index, timestamp))
-    #     self._eventloop.run_until_complete(task)
-    #     # log.debug(f"Added job: {self._jobs[index][0].args}")
-    #
-    # async def _async_add_job(self, job, index, timestamp):
-    #     encoded_job = self._encode_job(job)
-    #     job_tuple = tuple([encoded_job, timestamp])
-    #     async with self.config.jobs() as jobs:
-    #         jobs.insert(index, job_tuple)
-    #     # await self.config.jobs_index.set_raw(job.id, value=job_tuple)
-    #     return True
-
-    # @run_in_event_loop
-    # def update_job(self, job):
-    #     old_tuple: Tuple[Union[Job, None], Union[datetime, None]] = self._jobs_index.get(
-    #         job.id, (None, None)
-    #     )
-    #     old_job = old_tuple[0]
-    #     old_timestamp = old_tuple[1]
-    #     if old_job is None:
-    #         raise JobLookupError(job.id)
-    #
-    #     # If the next run time has not changed, simply replace the job in its present index.
-    #     # Otherwise, reinsert the job to the list to preserve the ordering.
-    #     old_index = self._get_job_index(old_timestamp, old_job.id)
-    #     new_timestamp = datetime_to_utc_timestamp(job.next_run_time)
-    #     task = asyncio.create_task(
-    #         self._async_update_job(job, new_timestamp, old_index, old_job, old_timestamp)
-    #     )
-    #     self._eventloop.run_until_complete(task)
-    #
-    # async def _async_update_job(self, job, new_timestamp, old_index, old_job, old_timestamp):
-    #     encoded_job = self._encode_job(job)
-    #     if old_timestamp == new_timestamp:
-    #         self._jobs[old_index] = (job, new_timestamp)
-    #         async with self.config.jobs() as jobs:
-    #             jobs[old_index] = (encoded_job, new_timestamp)
-    #     else:
-    #         del self._jobs[old_index]
-    #         new_index = self._get_job_index(new_timestamp, job.id)  # This is fine
-    #         self._jobs.insert(new_index, (job, new_timestamp))
-    #         async with self.config.jobs() as jobs:
-    #             del jobs[old_index]
-    #             jobs.insert(new_index, (encoded_job, new_timestamp))
-    #     self._jobs_index[old_job.id] = (job, new_timestamp)
-    #     # await self.config.jobs_index.set_raw(old_job.id, value=(encoded_job, new_timestamp))
-    #
-    #     log.debug(f"Async Updated {job.id=}")
-    #     # log.debug(f"Check job args: {job.kwargs=}")
-
-    # @run_in_event_loop
-    # def remove_job(self, job_id):
-    #     """Copied instead of super for the asyncio args"""
-    #     job, timestamp = self._jobs_index.get(job_id, (None, None))
-    #     if job is None:
-    #         raise JobLookupError(job_id)
-    #
-    #     index = self._get_job_index(timestamp, job_id)
-    #     del self._jobs[index]
-    #     del self._jobs_index[job.id]
-    #     task = asyncio.create_task(self._async_remove_job(index, job))
-    #     self._eventloop.run_until_complete(task)
-    #
-    # async def _async_remove_job(self, index, job):
-    #     async with self.config.jobs() as jobs:
-    #         del jobs[index]
-    #     # await self.config.jobs_index.clear_raw(job.id)
-
     @run_in_event_loop
     def remove_all_jobs(self):
         super().remove_all_jobs()
@@ -201,4 +124,5 @@ class RedConfigJobStore(MemoryJobStore):
 
     async def async_shutdown(self):
         await self.save_to_config()
-        super().remove_all_jobs()
+        self._jobs = []
+        self._jobs_index = {}

fifo/task.py

@@ -1,18 +1,19 @@
 import logging
 from datetime import datetime, timedelta
-from typing import Dict, List, Union
+from typing import Dict, List, Optional, Tuple, Union
 
 import discord
 from apscheduler.triggers.base import BaseTrigger
 from apscheduler.triggers.combining import OrTrigger
 from apscheduler.triggers.cron import CronTrigger
-from apscheduler.triggers.date import DateTrigger
 from apscheduler.triggers.interval import IntervalTrigger
 from discord.utils import time_snowflake
-from pytz import timezone
+import pytz
 from redbot.core import Config, commands
 from redbot.core.bot import Red
 
+from fifo.date_trigger import CustomDateTrigger
+
 log = logging.getLogger("red.fox_v3.fifo.task")
 
 
@@ -26,7 +27,7 @@ def get_trigger(data):
         return IntervalTrigger(days=parsed_time.days, seconds=parsed_time.seconds)
 
     if data["type"] == "date":
-        return DateTrigger(data["time_data"], timezone=data["tzinfo"])
+        return CustomDateTrigger(data["time_data"], timezone=data["tzinfo"])
 
     if data["type"] == "cron":
         return CronTrigger.from_crontab(data["time_data"], timezone=data["tzinfo"])
@@ -34,14 +35,25 @@ def get_trigger(data):
     return False
 
 
+def check_expired_trigger(trigger: BaseTrigger):
+    return trigger.get_next_fire_time(None, datetime.now(pytz.utc)) is None
+
+
 def parse_triggers(data: Union[Dict, None]):
     if data is None or not data.get("triggers", False):  # No triggers
         return None
 
     if len(data["triggers"]) > 1:  # Multiple triggers
-        return OrTrigger([get_trigger(t_data) for t_data in data["triggers"]])
+        triggers_list = [get_trigger(t_data) for t_data in data["triggers"]]
+        triggers_list = [t for t in triggers_list if not check_expired_trigger(t)]
+        if not triggers_list:
+            return None
+        return OrTrigger(triggers_list)
     else:
-        return get_trigger(data["triggers"][0])
+        trigger = get_trigger(data["triggers"][0])
+        if check_expired_trigger(trigger):
+            return None
+        return trigger
 
 
 class FakeMessage:
@@ -66,11 +78,11 @@ def neuter_message(message: FakeMessage):
 
 
 class Task:
-    default_task_data = {"triggers": [], "command_str": ""}
+    default_task_data = {"triggers": [], "command_str": "", "expired_triggers": []}
 
     default_trigger = {
         "type": "",
-        "time_data": None,  # Used for Interval and Date Triggers
+        "time_data": None,
         "tzinfo": None,
     }
 
@@ -87,9 +99,10 @@ class Task:
 
     async def _encode_time_triggers(self):
         if not self.data or not self.data.get("triggers", None):
-            return []
+            return [], []
 
         triggers = []
+        expired_triggers = []
         for t in self.data["triggers"]:
             if t["type"] == "interval":  # Convert into timedelta
                 td: timedelta = t["time_data"]
@@ -101,13 +114,15 @@ class Task:
 
             if t["type"] == "date":  # Convert into datetime
                 dt: datetime = t["time_data"]
-                triggers.append(
+                data_to_append = {
-                    {
+                    "type": t["type"],
-                        "type": t["type"],
+                    "time_data": dt.isoformat(),
-                        "time_data": dt.isoformat(),
+                    "tzinfo": getattr(t["tzinfo"], "zone", None),
-                        "tzinfo": getattr(t["tzinfo"], "zone", None),
+                }
-                    }
+                if dt < datetime.now(pytz.utc):
-                )
+                    expired_triggers.append(data_to_append)
+                else:
+                    triggers.append(data_to_append)
                 continue
 
             if t["type"] == "cron":
@@ -125,7 +140,7 @@ class Task:
 
             raise NotImplemented
 
-        return triggers
+        return triggers, expired_triggers
 
     async def _decode_time_triggers(self):
         if not self.data or not self.data.get("triggers", None):
@@ -138,7 +153,7 @@ class Task:
 
             # First decode timezone if there is one
             if t["tzinfo"] is not None:
-                t["tzinfo"] = timezone(t["tzinfo"])
+                t["tzinfo"] = pytz.timezone(t["tzinfo"])
 
             if t["type"] == "interval":  # Convert into timedelta
                 t["time_data"] = timedelta(**t["time_data"])
@@ -174,14 +189,23 @@ class Task:
         await self._decode_time_triggers()
         return self.data
 
-    async def get_triggers(self) -> List[Union[IntervalTrigger, DateTrigger]]:
+    async def get_triggers(self) -> Tuple[List[BaseTrigger], List[BaseTrigger]]:
         if not self.data:
             await self.load_from_config()
 
         if self.data is None or "triggers" not in self.data:  # No triggers
-            return []
+            return [], []
+
+        trigs = []
+        expired_trigs = []
+        for t in self.data["triggers"]:
+            trig = get_trigger(t)
+            if check_expired_trigger(trig):
+                expired_trigs.append(t)
+            else:
+                trigs.append(t)
 
-        return [get_trigger(t) for t in self.data["triggers"]]
+        return trigs, expired_trigs
 
     async def get_combined_trigger(self) -> Union[BaseTrigger, None]:
         if not self.data:
@@ -201,7 +225,10 @@ class Task:
         data_to_save = self.default_task_data.copy()
         if self.data:
             data_to_save["command_str"] = self.get_command_str()
-            data_to_save["triggers"] = await self._encode_time_triggers()
+            (
+                data_to_save["triggers"],
+                data_to_save["expired_triggers"],
+            ) = await self._encode_time_triggers()
 
         to_save = {
             "guild_id": self.guild_id,
@@ -217,7 +244,10 @@ class Task:
             return
 
         data_to_save = self.data.copy()
-        data_to_save["triggers"] = await self._encode_time_triggers()
+        (
+            data_to_save["triggers"],
+            data_to_save["expired_triggers"],
+        ) = await self._encode_time_triggers()
 
         await self.config.guild_from_id(self.guild_id).tasks.set_raw(
             self.name, "data", value=data_to_save
@@ -247,12 +277,16 @@ class Task:
             )
             return False
 
-        actual_message: discord.Message = channel.last_message
+        actual_message: Optional[discord.Message] = channel.last_message
         # I'd like to present you my chain of increasingly desperate message fetching attempts
         if actual_message is None:
             # log.warning("No message found in channel cache yet, skipping execution")
             # return
-            actual_message = await channel.fetch_message(channel.last_message_id)
+            if channel.last_message_id is not None:
+                try:
+                    actual_message = await channel.fetch_message(channel.last_message_id)
+                except discord.NotFound:
+                    actual_message = None
             if actual_message is None:  # last_message_id was an invalid message I guess
                 actual_message = await channel.history(limit=1).flatten()
                 if not actual_message:  # Basically only happens if the channel has no messages