Merge pull request #81 from foarsitter/recipes

Explain how common use cases are handled

docs/index.md

@@ -15,6 +15,7 @@ maxdepth: 1
 ---
 
 usage
+recipes
 reference
 contributing
 Code of Conduct <codeofconduct>

docs/recipes.md

@@ -0,0 +1,107 @@
+from requestmodel.typing import ResponseType
+
+# Recipes
+
+This document outlines some special cases you will encounter in the wild.
+
+## Headers with dashes
+
+In Python, and most other languages, it is not possible to use dashes in variable names. However, in HTTP headers,
+dashes
+are used by convention.
+
+This can be handled in two ways
+
+1. `convert_underscores` which defaults to true will convert underscores to dashes in the headers. This is default
+   behavior so you will need to set it to false if you want to keep the underscores.
+2. Another escape hatch is to use the `alias` keyword.
+
+## Embed keys in request body
+
+By design `RequestModel` will ignore the outer key in the request body. This is done to map a pydantic `BaseModel` to
+the top level
+of the request body. If you want to include the outer key, you can use the `embed` variable in the Body parameter.
+
+```python
+from typing import Annotated
+
+from pydantic import BaseModel
+from requestmodel import RequestModel
+from requestmodel.params import Body
+
+
+class PersonForm(BaseModel):
+    name: str
+    age: int
+
+
+class PersonFormRequest(RequestModel[PersonForm]):
+    method = "POST"
+    url = "/api/v1/person/create"
+    # a BaseModel will be added to the body by default
+    body: PersonForm
+    # so this is the same as above
+    body: Annotated[PersonForm, Body()]
+    # remember to set the response_model to the object instead of annotating its type
+    response_model = PersonForm
+```
+
+In the above example the name and age paremeters will be sent as the top level of the request body.
+
+```json
+{
+  "name": "John",
+  "age": 42
+}
+```
+
+If you want to include the outer key, you can use the `embed` variable in the Body parameter.
+
+```python
+    body: Annotated[Form, Body(embed=True)]
+```
+
+Age and name will not be nested in the body key
+
+```json
+{
+  "body": {
+    "name": "John",
+    "age": 42
+  }
+}
+```
+
+## Python types as request and responses
+
+Most API's will return a dictionary to embed a list response behind a key to provide more meta information, in example for providing information about the pagination object.
+
+However, sometimes you will encounter API's that return a list directly. To enable serialization of these responses you need to make use of a TypeAdapter provided by pydantic.
+
+```python
+from pydantic import BaseModel
+from pydantic import TypeAdapter
+from requestmodel import IteratorRequestModel
+
+class Person(BaseModel):
+    name: str
+    age: int
+
+
+PersonList = TypeAdapter[list[Person]]
+
+# as the generic type, use the plain python type
+class PersonRequest(IteratorRequestModel[list[Person]]):
+    method = "GET"
+    url = "/api/v1/persons"
+    limit: int = 10
+    offset: int = 0
+
+    # to make serialization possible, use the TypeAdapter as the response_model
+    response_model = PersonList
+
+    def next_from_response(self, response: list[Person]) -> bool:
+        self.offset += self.limit
+        return len(response) >= self.limit
+
+```

docs/usage.md

@@ -20,7 +20,7 @@ class MyResponse(BaseModel):
 # the goal is to follow the rules FastAPI uses to describe endpoints
 class MyRequest(RequestModel[MyResponse]):
     method = "GET"
-    url = "https://example.com/api/v1/my-endpoint/{param1}"
+    url = "/api/v1/my-endpoint/{param1}"
 
     param1: str  # will be used in the url
     param2: int  # will be converted as a query parameter
@@ -33,7 +33,7 @@ class MyRequest(RequestModel[MyResponse]):
     response_model = MyResponse
 
 
-client = Client()
+client = Client(base_url="https://example.com")
 
 response = MyRequest(param1="foo", param2=42).send(client)