#django

mar77i: https://docs.djangoproject.com/en/1.10/ref/cont...

prsn: write a template tag then. but to be honest, i'd just edit the template, probably.

If I want to link to this in reStructured Text how would I do it? https://docs.djangoproject.com/en/1.10/ref/requ...

this is what I tried: :class`~django.http.HttpRequest`'s :ref:`META<django.http.HttpRequest.META>`

Ryanar: link from where? from inside django's docs?

yeah

the page has the following label .. _httprequest-attributes:

Ryanar: i can never remember the details of the rst / sphinx syntax. i'd look up an equivalent that is already there and copy/paste/edit it :p

mar77i: User.save wouldn't know if .password is already encrypted or not

got it, :class:`~django.http.HttpRequest.META`

hello

let's say I have two unrelated models, each with an email field

I want to join these models,

can the ORM help me?

typically I'd like to select instances of one model, filtered by the existence of a related model matching some conditions

I'm aware of the possible performance issues of such a join. just curious if there's a way to do it without dropping to raw SQL (which is a fine answer there)

apollo13: fooling around... err... A/B testing shows that it's related to the django-cms placeholder cache :-/

mYk - Model1.objects.filter(email=Model2.objects.get(email=email))?

you can do that?

well I don't think this is going to work very well

shouldn't it be email__in=...?

yeah that answer made me realize I can do it with a subquery

mYk: iirc you can't do a db-level join in that case

s/I can/perhaps I can/

Model1.objects.filter(email__in=Model2.objects.filter(...).values_list('email', flat=True)) -- not sure that works, and if it does, it's going to fetch the list and pass it back to a second query

mYk: something like that should work, yep

for a real join, you need raw sql, afaik

yeah that's the conclusion I'm reaching as well

thanks for your help!

mYk: you might be able to drop .values_list, django should be smart enough to figure it out (tm)

needs testing, then

at least if it is the pk, if not, you can drop flat at least

so probably just .values('email') if email != pk

I'm still at the thought experiment stage because I'm writing a spec

oh?

spec for what :)

spamming

err

nagging users who requested a quote until they purchase a plan

:^)

it's a bit complicated since they may have reached any stage of our 4-step subscription pipeline an arbitrary number of times, and each of these may trigger different emails, but you don't want to send *all* these emails

hence the not-entirely-trivial design

so you might be able to make use of union and intersection :D

apollo13: https://dpaste.de/8bmC do you happen to have any idea?

mYk - you can use union

moldy: well, pickle works recursive, so you have to drill down through the response

moldy: also check if response.request or so exists

that said, I think pickling responses is buggy at best

apollo13: hmm, ok, thanks

but the page cache relies on this, right?

moldy: basically do dir(value) and then if request is there do del response.request and see if it pickles

if yes the request is at fault :D

yeah

page cache pickles the http response, but we had a few errors like that in the past

imo the page cache should pickle body + headers, not the whole response

moldy: oh I was wrong, try if value._closable_objects.clear() and then dumping fixes it

apollo13: nope, same problem

apollo13: i should add that type(value) is <class 'rest_framework.response.Response'>, might well be related to that

stupid questio, but is the response already rendered?

oh, mhm

if that adds weird extra thingies, then for sure

moldy: https://github.com/tomchristie/django-rest-fram... :D

moldy: you could run response.__getstate__ and see what comes out

stuff ;p

obviously, but copy that into an editor and search for lambda?

maybe throw through prettyprint?

apollo13: yeah, good idea

hi, i want to know if i define explicitly in database router to route all read/writes to the 'default' db, and then use Mode.objects.using('another_db').all() , that will overrides the router and hits `another_db`?

You don't even need a router for that. By default, reads & writes will already use the default DB.

@mattmcc : yes, but i wanted to make sure no read/write goes to any other db by mistake

because have same models on both db s

i also needed router to prevent `another_db` from migration

I can shell in and call_command a management command of mine, with the appropriate settings file, etc. Works great. When I try to do so on the command line, python manage.py my_command... I get all these weird import errors. Ive run out of ideas of things to consider.. Same settings file, same python, same virtualenv.

pannonia: can you show us the errors?

ah thanks well it'd take a while to clean out sensitive stuff, but the errors are all ImportErrors. The management command imports a model that imports something else etc... Again works fine as django.core.management.call_command but wont work as straight python call...

I've given full path to .virtualenv/project/bin/python and full path to manage.py. Just running out of things to check

without seeing any error, we don't have much basis for guesses

circular imports?

yeah. hm. ok well Ill spend some time recreating issue in simpler code I can share. Appreciate your taking a stab at it.