3) Usually you don't need to close the connections explicitly - they are returned to the pool after each request handling. But in case when your handling code would take a long time processing e.g. I/O, you may want to turn off "use_connection_for_request" (similar docs: https://python-gino.org/docs/en/1.0/reference/extensions/starlette.html#lazy-connection) or explicitly return the connection.
my following question would be, when did your code tried to access database and ended up with
UninitializedError?
is it in a request handler, or somewhere at the top level of your
main.py?
from sanic import Sanic
from gino.ext.sanic import Gino
app = Sanic("app name")
app.config.DB_HOST = "127.0.0.1"
app.config.DB_USER = "fantixking"
db = Gino()
db.init_app(app)
class InfoModel(db.Model):
__tablename__ = "infomodel"
@app.route("/test")
async def test(request):
info = await InfoModel.query.gino.first()
ret = 'INFO MODEL domain: %s' % info.domain
return HTTPResponse(ret, content_type='text/html')
if __name__ == "__main__":
app.run(host="0.0.0.0", port=8000)
this works for me
was the request sent too soon after server start up? I know uvicorn had similar problem but it's fixed now
I'm starting the server using:
python -m MyServer.main
I also have Python 3.6.9
Here's the full traceback
Traceback (most recent call last):
File "####/env/lib/python3.6/site-packages/sanic/app.py", line 976, in handle_request
response = await response
File "/usr/lib/python3.6/asyncio/coroutines.py", line 110, in __next__
return self.gen.send(None)
File "####/project/handlers.py", line 12, in test
info = await InfoModel.query.gino.first()
File "/usr/lib/python3.6/asyncio/coroutines.py", line 110, in __next__
return self.gen.send(None)
File "####/env/lib/python3.6/site-packages/gino/api.py", line 137, in first
return await self._query.bind.first(self._query, *multiparams, **params)
File "####/env/lib/python3.6/site-packages/gino/api.py", line 544, in __getattribute__
raise self._exception
gino.exceptions.UninitializedError: Gino engine is not initialized.This issue has something to do with the files being split up. I used your example and incrementally added my configs, and whatnot. It started going wrong as soon as I put the handler in it's own file.
I have 2 identical (save for the path) handlers: one on the test.py file and one in a module's handlers.py file. The one on test.py works fine.
@fantix @wwwjfy
I'm trying to use pool_pre_ping argument as shown in the docs here, but it's throwing the following error. https://docs.sqlalchemy.org/en/13/core/pooling.html#disconnect-handling-pessimistic
Any help here would be appreciated. Thanks :)
I'm trying to use pool_pre_ping argument as shown in the docs here, but it's throwing the following error. https://docs.sqlalchemy.org/en/13/core/pooling.html#disconnect-handling-pessimistic
'pool_pre_ping' sent to create_engine(), using configuration AsyncpgDialect/GinoEngine. Please checkthat the keyword arguments are appropriate for this combination of components.Any help here would be appreciated. Thanks :)
2 replies
Hi chat. I'm stuck with a usecase of mine corresponding the loaders.
So I've written down my own "ON" clause as the tables in consideration don't have and foreign-key relation. But I'm not sure how to load distinct children in the Parent using the "distinct" keyword.
This is loading the same Child instance in the parent again-and-again and I'm not sure how to use the distinct keyword here.
So I've written down my own "ON" clause as the tables in consideration don't have and foreign-key relation. But I'm not sure how to load distinct children in the Parent using the "distinct" keyword.
loader = Parent.load(add_child=Childern.on(child_and_clause))result = (
await loader.query.where(Parent.id == parent_id)
.where(Parent.knowledge_end_date == None)
.gino.first()
)This is loading the same Child instance in the parent again-and-again and I'm not sure how to use the distinct keyword here.
child_and_clause is working fine as the output is correct, justa that it is being added multiple times
I am using the sanic ext:
from gino.ext.sanic import Gino
db = Gino()
I'm using Gino inside my Celery apps using this line
async with db.with_bind(settings.SQL_DRIVER, echo=False) as engine:
For some reason I can't get echo to turn off.
Hi! @rlundy82. I came across the same issue while using uvicorn with Sanic+Gino. the easiest way to disable db echo is to set a higher logging level to gino_logger:
import logging
logging.getLogger("gino").setLevel(logging.WARNING)
@fantix @wwwjfy Do we have sqlalchemy type events in place? (like after_create, after_update, etc.) Or is there any decent way to achieve this?
Solution provided here looks a bit hacky and I'm not very convinced if I should implement something like this (If any of you approves, I can go and implement a similar setup). Thanks :)
Solution provided here looks a bit hacky and I'm not very convinced if I should implement something like this (If any of you approves, I can go and implement a similar setup). Thanks :)
then call
create_user() instead of User.create()
or alternatively override
User.create method
Just my first reaction to this question, would require more use cases though
such hooks could be added
but the
create() and create_all() functions are a bit less elegant - they're copied from SQLAlchemy and manipulated to be async, so I'm overall not a big fan
the
create_async or GinoSchemaVisitor.create might be the places to override
Hi guys, does anybody knows how to wrap all pytest (pytest-asyncio) + sanic/aiohttp test cases in transactions? Seems like there is multiple different event loops, and asyncpg can't work in this conditions (similar issue like I have: python-gino/gino#713)
1 reply
As alternative solution I'm already do TRUNCATE at end of each test case, but it is too slow, nearly 1 second for each test case. Transactions can improve mine test run time significally, I think.
As final solution I want have something like django test cases, where transaction starts at setup_method and rollbacks at teardown_method.