Issue with sqlalchemy (cannot join current thread and working outside of app context)

[stautonico]

Hi, I'm trying to create a background job that interacts with my database via flask-sqlalchemy and am having some trouble. Here is some (simplified) code that I'm trying to run:

# scheduler.py
from flask_apscheduler import APScheduler
scheduler = APScheduler()

# app.py
from myapp.scheduler import scheduler

app = Flask(__name__)
# ... setup stuff
scheduler.init_app(app)

# more setup stuff
configure_scheduler(app)



def configure_scheduler(app):
    def delete_expired_shares():
        with app.app_context():
            shares = Shares.query.filter(Shares.expires <= dt.now())

    scheduler.add_job(func=delete_expired_shares, trigger="interval", seconds=5, id="delete-expires-shares")
    scheduler.start()

When I include the with app.app_context(), I get the error: RuntimeError: cannot join current thread, but if I remove the with app.app_context(), it obv doesn't work (Working outside of application context). Is there something I missed in the docs? I looked at the examples and I think I set everything up correctly.

[jneeee]

Hi, Check if sqlalchemy have a config like check_same_thread=False?

        self.connection = sqlite3.connect(dbpath, check_same_thread=False)

[stautonico]

I'm not quite sure what happened, but it seems to work now? I literally changed nothing. I just haven't touched the project since the 7th. I remember having to use check_same_thread in another unrelated program. I set the option and it still continues to work. I'm not sure what to think of this, but ig this is resolved?

[stautonico]

I think I found the source of my issue. I tried running a job that does a db.session.commit(), and that seems to fail at the end of the job. I put a print statement at the very end, which prints right before the program throws an exception. I tried setting SQLALCHEMY_CHECK_SAME_THREAD in my app.config as well as setting ?check_same_thread=False at the end of my sqlite URI, none of which solved my issue.

[stautonico]

I noticed that when I comment out my @app.teardown_appcontext, it seems to work. I think I might be using teardown appcontext wrong:

@app.teardown_appcontext
def stop_scheduler(exception=None):
    scheduler.shutdown()