genericclient-aiohttp 1.4.0

Instantiation

myclient = GenericClient(url, auth=None, session=None, trailing_slash=False, retries=3, autopaginate=None)

Arguments:

• url: The root URL of your API

• auth: The auth for your API. You can pass anything that aiohttp.ClientSession can accept as auth.

• session: Pass a session instance to have aiohttp use that session. If None (the default), it will instantiate an instance of aiohttp.ClientSession for you.

• trailing_slash: You can set this to True if your API’s URLs end with a /

• retries: How many times should the client retry the http call after a ClientConnectionError

• autopaginate: You can set this to a coroutine to fetch all pages resulting from a request. Currently, the only coroutine included is genericclient_aiohttp.pagination.link_header, which supports [RFC5988](https://tools.ietf.org/html/rfc5988).

Endpoints

Endpoints are available as properties or items on the main instance:

await myclient.posts.all()  # GET /posts/
await myclient["active-users"].all()  # GET /active-users/

await myclient.posts.all()  # GET /posts/

await myclient.posts.filter(blog=12, status=1)  # GET /posts/?blog=12&status=1

await myclient.posts.filter(blog=12, status=1)  # GET /posts/?blog=12&status=1
await myclient.posts.filter(id=12)  # GET /posts/12/
await myclient.posts.filter(slug='12-ways-clickbait')  # GET /posts/12-ways-clickbait/

post = await myclient.posts.create({'blog': 12, 'status': 1})  # POST /posts/

# Assuming it doesn't exist
post = await myclient.posts.get_or_update(slug='my-post', defaults={'status': 1})  # GET /posts/my-post/, then POST /posts/

post = await myclient.posts.create_or_update({'status': 1})  # POST /posts/
post = await myclient.posts.create_or_update({'id': 1234, 'status': 1})  # PUT /posts/1234/

post = await myclient.posts.create_or_update({'id': 1234})  # PUT /posts/1234/
# <- server returns 400
# -> PATCH /posts/1234/

await myclient.posts.delete(24)  # DELETE /posts/24/

Resources

All endpoints methods (with the exception of .delete()) return either a Resource or a list of Resource s.

A Resource is just a wrapping class for a dict, where keys can be accessed as properties.

Additionally, Resource s have a special property called .payload, which contains the original payload received from the server.

Resource s have the following methods:

Resource.delete() will result in a DELETE, with Resource.id as par of the URL:

blog = await myclient.posts.create({'blog': 12, 'status': 1})  # POST /posts/
await blog.delete()  # DELETE /blog/345/ -- the ID 345 was returned by the server in the previous response

Resource.save() will result in a PUT, with Resource.id as par of the URL. If the server returns a 400 error, a PATCH request will be re-issued:

post = await myclient.posts.create({'blog': 12, 'status': 1})  # POST /posts/
post.status = 2
await post.save()  # PUT /posts/345/

post = Resource(id=345, status=1)
await post.save()  # PUT /posts/345/
# <- server returns 400
# -> PATCH /posts/345/

ResourceSets

Whenever a method returns a list of Resources, they list will be wrapped in a ResultSet.

A ResultSet is a just a list object, with the addition of a .response containing the original response from the server.

Customizing Endpoints and Resources

Resources can be customized by subclassing genericclient_aiohttp.Resource.

The most common reason is specifying the name of the primary key:

from genericclient_aiohttp import Resource


class PostResource(Resource):
    pk_name = 'slug'

Endpoints can be customized by subclassing genericclient_aiohttp.Endpoint:

form genericclient_aiohttp import Endpoint


class PostEndpoint(Endpoint):
    resource_class = PostResource

You can then subclass genericclient_aiohttp.GenericClient to tell the client which endpoint classes to use on each endpoint:

from genericclient_aiohttp import GenericClient

class Client(GenericClient):
    endpoint_classes = {
        'posts': PostEndpoint,
    }

Routes

If your API has some non-RESTful calls within the main endpoints (sometimes referred as detail_route and list_route), you can use genericclient to call them:

await myclient.posts(id=123).publish(date=tomorrow)

await myclient.blogs().ping()

Routes http calls use POST by default, but you can specify something else by using the _method argument:

await myclient.posts(_method='get', id=123).pingbacks()

await myclient.blogs(_method='get').visits()

Note that this calls will return an instance of genericclient.ParsedResponse, instead of instances of genericclient.Resource,