Add option to provide a callable for PolymorphicProxySerializer.serializers #953
Conversation
Codecov ReportPatch coverage:
Additional details and impacted files@@ Coverage Diff @@
## master #953 +/- ##
==========================================
+ Coverage 98.61% 98.63% +0.01%
==========================================
Files 68 68
Lines 8074 8110 +36
==========================================
+ Hits 7962 7999 +37
+ Misses 112 111 -1
Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here. ☔ View full report at Codecov. |
|
@glennmatthews, awesome stuff! That make it even more meta 😆, I like it! Just one thought, do you think providing arguments is superfluous? (e.g. Not sure if this adds something, you have seen this pattern recurring. What do you think? Just a question, not a change request. |
|
Whatever, let's just merge it. We have a method to tack on optional kwargs to this callback. Still interested in your opinion though. drf-spectacular/drf_spectacular/plumbing.py Line 1338 in e98b0c6 |
glennmatthews
deleted the
u/glennmatthews-polymorphicproxyserializer-enhancements
branch
|
Thus far I haven't seen any need to pass args into the callable, but I like the way you're thinking. :-) Thanks for the approval and merge! |
Adding a minor feature that we found useful in our own use of
PolymorphicProxySerializer- the ability to initialize it with a callable (such as a lambda) that returns a list/dict ofserializerswhen evaluated, rather than needing to provide the list of serializers up front at declaration time.In our case this was useful because we have a polymorphic schema method field that can return a serializer corresponding to any subclass of a given abstract base model, not all of which may already be defined/imported at the time the field is declared. Being able to provide a callback that can be deferred until the time the schema is actually generated made our implementation easier.