Persist job schedules

The default memory store used by Quart-Tasks is wiped every time the app restarts. This is problematic if a task’s invocation is missed during a restart - as there is no way for Quart-Tasks to know and hence it skips the invocation.

It is possible to change the store to one that persists invocation times. If a task invocation is missed the store will return the previous invocation thereby prompting a new invocation straight after startup.

An example of a persisting store is shown below using Quart-DB, note it is assumed the database has a table called jobs with the columns name and executed with a unique constraint on the name column,

from quart import Quart
from quart_db import QuartDB
from quart_tasks import QuartTasks, TaskStoreABC

app = Quart(__name__)
quart_db = QuartDB(app)

class DBStore(TaskStoreABC):
    def __init__(self) -> None:
        pass

    async def startup(self) -> None:
        pass

    async def get(self, key: str, default: datetime) -> datetime:
        async with quart_db.connection() as connection:
            result = await connection.fetch_val(
                "SELECT executed FROM jobs WHERE name = :name",
                {"name": key},
            )
            return default if result is None else result

    async def set(self, key: str, executed: datetime) -> None:
        async with quart_db.connection() as connection:
            await connection.execute(
                """INSERT INTO jobs (name, executed)
                        VALUES (:name, :executed)
                   ON CONFLICT name DO UPDATE
                           SET executed = :executed""",
                {"name": key, "executed": executed},
            )

    async def shutdown(self) -> None:
        pass

 quart_tasks = QuartTasks(app, store=DBStore())