Clojure HTTP Client
by Dan Larkin and Phil Hagelberg
A work in progress.
There are two namespaces, clojure.http.client, which provides a simple
"request" function, and clojure.http.resourcefully, which is targeted
more towards interactions with REST-based APIs.
(ns clojure.http.example
(:use [clojure.http.client]
[clojure.contrib.json.write])
(:require [clojure.http.resourcefully :as resourcefully]))
(let [response (request "http://google.com")]
(:code response) ;; 200
(:msg response) ;; "OK"
(:body-seq response)) ;; ("<html><head><meta[...]" ...
(resourcefully/put "http://localhost:5984/my-db/doc1"
{} (json-str {:hello "world"}))
(resourcefully/with-cookies {}
(resourcefully/post "http://localhost:3000/login"
{} {"user" user "password" password})
(resourcefully/get "http://localhost:3000/my-secret-page))
The request function requires a URL and optionally accepts a method
(GET by default), a map of headers, a map of cookies, and a request
body. The resourcefully functions take a URL, an optional headers map,
and an optional body.
Request bodies may be strings, maps, or InputStreams. Strings get sent
verbatim. Maps get sent as application/x-www-form-urlencoded, and
InputStreams get streamed to the server 1000 bytes at a time.
The functions in resourcefully are named after the HTTP verbs. Note
that resourcefully must be required :as something since it defines a
"get" function, which would interfere with core if it were fully
referred. Exceptions will be raised for status codes that indicate
problems, so you don't have to check return codes manually. If you use
resourcefully inside a "with-cookies" block, cookies will
automatically be saved in a cookies ref and sent out with each
request.
TODO:
- Connection pooling/keep-alive?
- Anything else? Send a message via github or the Clojure mailing list.
Licensed under the same terms as Clojure.