Merge pull request #77 from chanzuckerberg/ninabernick/allow_custom_q…

…ueries_mutations

Custom root level queries and mutations via inheritance

docs/custom_code.md

@@ -0,0 +1,53 @@
+# How To: Extending Platformics Generated Code
+Platformics will generate a GraphQL API that exposes CRUD operations for each entity type in your LinkML schema. It is often the case, however, that you will need to expose more functionality in your API than just CRUD. There are a few options for extending generated code to add custom business logic to your application.
+
+## Adding root-level queries and mutations
+Platformics will generate basic read and aggregate queries in `api/queries.py`. These are inherited by a query class in `custom_queries.py`, which will gain access to all the codegenned queries as well as allow you to implement your own queries. To add a new root level field (query), add the field to the query class in `custom_queries.py`:
+
+```python
+@strawberry.type
+class Query(BaseQuery):
+    foo: str = my_custom_field
+```
+The resolver for the custom field should be decorated with `@strawberry.field`:
+```python
+@strawberry.field
+def my_custom_field(self) -> str:
+    return "foo"
+```
+
+Similarly, for custom mutations, root level mutations can be added to the `Mutation` class in `custom_mutations.py`, which inherits from the codegenned `Mutation` class in `api/mutations.py`
+
+```python
+@strawberry.type
+class Mutation(BaseMutation):
+    bar: str = my_custom_mutation
+```
+The resolver for the custom mutation should be decorated with `@strawberry.mutation`:
+```python
+@strawberry.mutation
+def my_custom_mutation(self) -> str:
+    return "bar"
+```
+Both the custom query and mutation classes are imported into `main.py` and used by Strawberry to generate an API at runtime:
+```python
+from custom_mutations import Mutation
+from custom_queries import Query
+
+schema = strawberry.Schema(query=Query, mutation=Mutation, config=get_strawberry_config(), extensions=[HandleErrors()])
+```
+
+### Overriding queries and mutations
+By creating a field with the same name as an existing codegenned query or mutation, you can override the codegenned resolver.
+```python
+@strawberry.mutation
+def my_override_mutation(self) -> str:
+    return "Sorry, I overrode the mutation"
+
+@strawberry.type
+class Mutation(BaseMutation):
+    create_sample: str = my_override_mutation
+```
+
+## Using your own templates
+TODO: fill this out. Existing functionality.

test_app/custom_mutations.py

@@ -0,0 +1,17 @@
+import strawberry
+from api.mutations import Mutation as BaseMutation
+
+
+@strawberry.mutation
+def my_custom_mutation(self) -> str:
+    return "bar"
+
+@strawberry.mutation
+def my_override_mutation(self) -> str:
+    return "Sorry, I overrode the mutation"
+
+
+@strawberry.type
+class Mutation(BaseMutation):
+    bar: str = my_custom_mutation
+    create_sample: str = my_override_mutation

test_app/custom_queries.py

@@ -0,0 +1,12 @@
+import strawberry
+from api.queries import Query as BaseQuery
+
+
+@strawberry.field
+def my_custom_field(self) -> str:
+    return "foo"
+
+
+@strawberry.type
+class Query(BaseQuery):
+    foo: str = my_custom_field

test_app/main.py

@@ -9,8 +9,9 @@
 from platformics.settings import APISettings
 from database import models
 
-from api.mutations import Mutation
-from api.queries import Query
+from custom_mutations import Mutation
+from custom_queries import Query
+
 
 settings = APISettings.model_validate({})  # Workaround for https://github.com/pydantic/pydantic/issues/3753
 schema = strawberry.Schema(query=Query, mutation=Mutation, config=get_strawberry_config(), extensions=[HandleErrors()])