Quickstart

Decorators and function annotations indicate how a request will be handled.

Request Method

Uplink offers decorators that turn any method into a request definition. These decorators provide the request method and relative URL of the intended request: get, post, put, patch and delete.

The relative URL of the resource is specified in the decorator.

@get("users/list")

You can also specify query parameters in the URL.

@get("users/list?sort=desc")

Moreover, request methods must be bound to a Consumer subclass.

class MyApi(Consumer):
    @get("users/list")
    def list_users(self):
        """List all users."""

URL Manipulation

A request URL can be updated dynamically using URI template parameters. A simple URI parameter is an alphanumeric string surrounded by { and }.

To match the parameter with a method argument, either match the argument’s name with the alphanumeric string, like so

@get("group/{id}/users")
def group_list(self, id): pass

or use the Path annotation.

@get("group/{id}/users")
def group_list(self, group_id: Path("id")): pass

Query parameters can also be added.

@get("group/{id}/users")
def group_list(self, group_id: Path("id"), sort: Query): pass

For complex query parameter combinations, a mapping can be used:

@get("group/{id}/users")
def group_list(self, group_id: Path("id"), options: QueryMap): pass

Request Body

An argument’s value can be specified for use as an HTTP request body with the Body annotation:

@post("users/new")
def create_user(self, user: Body): pass

This annotation works well with the keyword arguments parameter (denoted by the ** prefix):

@post("users/new")
def create_user(self, **user_info: Body): pass

Form Encoded, Multipart, and JSON

Methods can also be declared to send form-encoded, multipart, and JSON data.

Form-encoded data is sent when form_url_encoded decorates the method. Each key-value pair is annotated with a Field annotation:

@form_url_encoded
@post("user/edit")
def update_user(self, first_name: Field, last_name: Field): pass

Multipart requests are used when multipart decorates the method. Parts are declared using the Part annotation:

@multipart
@put("user/photo")
def update_user(self, photo: Part, description: Part): pass

JSON data is sent when json decorates the method. The Body annotation declares the JSON payload:

@uplink.json
@uplink.patch("/user")
def update_user(self, **user_info: uplink.Body):
    """Update an authenticated user."""

Header Manipulation

You can set static headers for a method using the headers decorator.

@headers({
    "Accept": "application/vnd.github.v3.full+json",
    "User-Agent": "Uplink-Sample-App"
})
@get("users/{username}")
def get_user(self, username): pass

headers can be used as a class decorator for headers that need to be added to every request:

@headers({
    "Accept": "application/vnd.github.v3.full+json",
    "User-Agent": "Uplink-Sample-App"
})
class GitHub(Consumer):
    ...

A request header can be updated dynamically using the Header function parameter annotation:

@get("user")
def get_user(self, authorization: Header):
    """Get an authenticated user."""

Synchronous vs. Asynchronous

By default, Uplink uses the Requests library to make requests. However, the client parameter of the Consumer constructor offers a way to swap out Requests with another HTTP client:

github = GitHub(BASE_URL, client=...)

Notably, Requests blocks while waiting for a response from a server. For non-blocking requests, Uplink comes with optional support for asyncio and twisted. Checkout this example on GitHub for more.

Deserializing the Response Body

Uplink makes it easy and optional to convert HTTP response bodies into data model objects, whether you leverage Uplink’s built-in support for libraries such as marshmallow (see uplink.converters.MarshmallowConverter) or use uplink.loads to write custom conversion logic that fits your unique needs.

At the least, you need to specify the expected return type using a decorator from the uplink.returns module. uplink.returns.json is handy when working with APIs that provide JSON responses:

@returns.json(User)
@get("users/{username}")
def get_user(self, username): pass

Python 3 users can alternatively use a return type hint:

@returns.json
@get("users/{username}")
def get_user(self, username) -> User: pass

The final step is to register a strategy that converts the HTTP response into the expected return type. To this end, uplink.loads() can register a function that handles such deserialization for a particular class and all its subclasses.

# The base class for all model types, including User from above.
from models import ModelBase

# Tell Uplink how to deserialize JSON responses into our model classes:
@loads.install  # Make this available to all consumer instances.
@loads.from_json(ModelBase)
def load_model_from_json(model_cls, json_obj):
    return model_cls.from_json(json_obj)

This step is not required if you define your data model objects using a library for whom Uplink has built-in support, such as marshmallow (see uplink.converters.MarshmallowConverter).

Note

For API endpoints that return collections (such as a list of users), Uplink just needs to know how to deserialize the element type (e.g., a user), offering built-in support for Converting Collections.

Custom Response and Error Handling

New in version 0.4.0.

To register a custom response or error handler, decorate a function with the response_handler or error_handler decorator.

Note

Unlike consumer methods, these functions should be defined outside of a class scope.

For instance, the function returns_success() defined below is a response handler that output whether or not the request was successful:

@uplink.response_handler
def returns_success(response):
    return response.status == 200

Now, returns_success() can be used as a decorator to inject its custom response handling into any request method:

@returns_success
@put("/todos")
def create_todo(self, title):
    """Creates a todo with the given title."""

To apply the function’s handling onto all request methods of a Consumer subclass, we can simply use the registered handler as a class decorator:

@returns_success
class TodoApp(uplink.Consumer):
    ...

Similarly, functions decorated with error_handler are registered error handlers. When applied to a request method, these handlers are invoked when the underlying HTTP client fails to execute a request:

@error_handler
def raise_api_error(exc_type, exc_val, exc_tb):
    # wrap client error with custom API error
    ...

Notably, handlers can be stacked on top of one another to chain their behavior:

@raise_api_error
@returns_success
class TodoApp(uplink.Consumer):
    ...

Annotating __init__() Arguments

New in version 0.4.0.

Function annotations like Query and Header can be used with constructor arguments of a Consumer subclass. When a new consumer instance is created, the value of these arguments are applied to all requests made through that instance.

For example, the following consumer accepts the API access token as the constructor argument access_token:

class GitHub(uplink.Consumer):

    def __init__(self, access_token: uplink.Query):
        ...

    @uplink.post("/user")
    def update_user(self, **info: Body):
        """Update the authenticated user"""

Now, all requests made from an instance of this consumer class will be authenticated with the access token passed in at initialization:

github = TodoApp("my-github-access-token")

# This request will include the above access token as a query parameter.
github.update_user(bio="Beam me up, Scotty!")

_inject() Request Properties

New in version 0.4.0.

As an alternative to Annotating __init__() Arguments, you can achieve a similar behavior with more control by using the Consumer._inject() method. With this method, you can calculate request properties within plain old python methods.

class TodoApp(uplink.Consumer):

    def __init__(self, username, password)
        # Create an access token
        api_key = create_api_key(username, password)

        # Inject it.
        self._inject(uplink.Query("api_key").with_value(api_key))

Similar to the annotation style, request properties added with _inject() method are applied to all requests made through the consumer instance.