[WIP] Add Query Optimizer that selects fields from query AST #56
Conversation
|
What do you think @zerolab?? |
NathHorrigan
force-pushed
the
feature/ast-query-optimizer
branch
from
5b78d91
to
b7a0d19
Compare
NathHorrigan
force-pushed
the
feature/ast-query-optimizer
branch
from
b7a0d19
to
641bcba
Compare
| @@ -178,3 +179,7 @@ | |||
| "ROUTING": "grapple.urls.channel_routing", | |||
| } | |||
| } | |||
|
|
|||
| # Query Optimisation helpers | |||
| SHELL_PLUS_PRINT_SQL = True | |||
There was a problem hiding this comment.
these should be enabled only locally, maybe?
| @@ -10,3 +10,4 @@ factory-boy==2.12.0 | |||
| wagtail-factories==2.0.0 | |||
| django-cors-headers==3.0.2 | |||
| wagtail-headless-preview==0.1.4 | |||
| django_extensions==2.2.9 | |||
There was a problem hiding this comment.
I am not convinced we need the extra dependency. Should only be a dev dependency
|
|
||
| if id is not None: | ||
| qs = qs.filter(pk=id) | ||
| qs = qs.get(id=id) |
There was a problem hiding this comment.
This can throw a DoesNotExist should the provided id not be there anymore
There was a problem hiding this comment.
Had a quick look and left a few comments.
Will try to review more in depth this weekend.
I am curious, do you have any before/after profiling stats?
incidentally, https://github.com/tolomea/django-auto-prefetch and https://github.com/jmcarp/nplusone popped up this week -- there some good ideas there
|
Why not consider using https://github.com/tfoxy/graphene-django-optimizer ? |
|
Hi @fabienheureux , I did try GDQ but it doesn't work with stream field and that's a big issue for us. Instead of hack GDQ into Grapple, I decided to create our own way of doing it that can be fine tuned and expanded in the future. |
# Conflicts: # example/home/models.py # grapple/urls.py
ruisaraiva19
changed the title
|
@NathHorrigan any chance you can bring this to the finish line? |
@NathHorrigan any news on this PR? I'm going to fix the merge conflicts 🙂 . |
|
I started a PR on the |
This is the initial code for optimizing GraphQL queries to reduce the number of database calls. The gist of this PR is that it analyses from the queries' AST what fields of what models it needs and then appends that to the Queryset object, further along the line when a
.specific()call is applied these field selects are called on the Queryset using.only(..),.select_related(..)&.prefetch_related(..).More accurate data will follow...