A client-side router for ClojureScript.
Add secretary to your project.clj
:dependencies
vector:
[clj-commons/secretary "1.2.4"]
For the current SNAPSHOT
version use:
[clj-commons/secretary "1.2.5-SNAPSHOT"]
To get started :require
secretary somewhere in your project.
(ns app.routes
(:require [secretary.core :as secretary :refer-macros [defroute]]))
Note: starting ClojureScript v0.0-2371, :refer
cannot be used to import macros into your project anymore. The proper way to do it is by using :refer-macros
as above. When using ClojureScript v0.0-2755 or above, if (:require [secretary.core :as secretary])
is used, macros will be automatically aliased to secretary
, e.g. secretary/defroute
.
Secretary is built around two main goals: creating route matchers and dispatching actions. Route matchers match and extract parameters from URI fragments and actions are functions which accept those parameters.
defroute
is Secretary's primary macro for defining a link between a
route matcher and an action. The signature of this macro is
[name? route destruct & body]
. We will skip the name?
part
of the signature for now and return to it when we discuss
named routes. To get clearer picture of this
let's define a route for users with an id.
(defroute "/users/:id" {:as params}
(js/console.log (str "User: " (:id params))))
In this example "/users/:id"
is route
, the route matcher,
{:as params}
is destruct
, the destructured parameters extracted
from the route matcher result, and the remaining
(js/console.log ...)
portion is body
, the route action.
Before going in to more detail let's try to dispatch our route.
(secretary/dispatch! "/users/gf3")
With any luck, when we refresh the page and view the console we should
see that User: gf3
has been logged somewhere.
By default the route matcher may either be a string or regular
expression. String route matchers have special syntax which may be
familiar to you if you've worked with Sinatra or
Ruby on Rails. When secretary/dispatch!
is called with a
URI it attempts to find a route match and it's corresponding
action. If the match is successful, parameters will be extracted from
the URI. For string route matchers these will be contained in a map;
for regular expressions a vector.
In the example above, the route matcher
"/users/:id"
successfully matched against "/users/gf3"
and
extracted {:id "gf3}
as parameters. You can refer to the table below
for more examples of route matchers and the parameters they return
when matched successfully.
Route matcher | URI | Parameters |
---|---|---|
"/:x/:y" |
"/foo/bar" |
{:x "foo" :y "bar"} |
"/:x/:x" |
"/foo/bar" |
{:x ["foo" "bar"]} |
"/files/*.:format" |
"/files/x.zip" |
{:* "x" :format "zip"} |
"*" |
"/any/thing" |
{:* "/any/thing"} |
"/*/*" |
"/n/e/thing" |
{:* ["n" "e/thing"]} |
"/*x/*y" |
"/n/e/thing" |
{:x "n" :y "e/thing"} |
#"/[a-z]+/\d+" |
"/foo/123" |
["/foo/123"] |
#"/([a-z]+)/(\d+)" |
"/foo/123" |
["foo" "123"] |
Now that we understand what happens during dispatch we can look at the
destruct
argument of defroute
. This part is literally sugar
around let
. Basically whenever one of our route matches is
successful and extracts parameters this is where we destructure
them. Under the hood, for example with our users route, this looks
something like the following.
(let [{:as params} {:id "gf3"}]
...)
Given this, it should be fairly easy to see that we could have have written
(defroute "/users/:id" {id :id}
(js/console.log (str "User: " id)))
and seen the same result. With string route matchers we can go even further and write
(defroute "/users/:id" [id]
(js/console.log (str "User: " id)))
which is essentially the same as saying {:keys [id]}
.
For regular expression route matchers we can only use vectors for destructuring since they only ever return vectors.
(defroute #"/users/(\d+)" [id]
(js/console.log (str "User: " id)))
If a URI contains a query string it will automatically be extracted to
:query-params
for string route matchers and to the last element for
regular expression matchers.
(defroute "/users/:id" [id query-params]
(js/console.log (str "User: " id))
(js/console.log (pr-str query-params)))
(defroute #"/users/(\d+)" [id {:keys [query-params]}]
(js/console.log (str "User: " id))
(js/console.log (pr-str query-params)))
;; In both instances...
(secretary/dispatch! "/users/10?action=delete")
;; ... will log
;; User: 10
;; "{:action \"delete\"}"
While route matching and dispatch is by itself useful, it is often
necessary to have functions which take a map of parameters and return
a URI. By passing an optional name to defroute
Secretary will
define this function for you.
(defroute users-path "/users" []
(js/console.log "Users path"))
(defroute user-path "/users/:id" [id]
(js/console.log (str "User " id "'s path"))
(users-path) ;; => "/users"
(user-path {:id 1}) ;; => "/users/1"
This also works with :query-params
.
(user-path {:id 1 :query-params {:action "delete"}})
;; => "/users/1?action=delete"
If the browser you're targeting does not support HTML5 history you can call
(secretary/set-config! :prefix "#")
to prefix generated URIs with a "#".
(user-path {:id 1})
;; => "#/users/1"
Beware that using :prefix
that way will make resulting URIs no longer compliant with standard URI syntax – the fragment must be the last part of the URI after the query). Indeed, the syntax of a URI is defined as:
scheme:[//[user:password@]host[:port]][/]path[?query][#fragment]
secretary
adds a #
after the path so it makes the fragment hide the query. For instance, the following URL is comprehended in different ways by secretary
and the spec:
https://www.example.com/path/of/app#path/inside/app?query=params&as=defined&by=secretary
the fragment is "path/inside/app?query=params&as=defined&by=secretary"
for standard libraries,
but is "path/inside/app"
according to Secretary
the query is ""
for standard libraries, but is "query=params&as=defined&by=secretary"
according to Secretary
You can extend Secretary's protocols to your own data types and records if you need special functionality.
IRenderRoute
Most of the time the defaults will be good enough but on occasion you
may need custom route rendering. To do this implement IRenderRoute
for your type or record.
(defrecord User [id]
secretary/IRenderRoute
(render-route [_]
(str "/users/" id))
(render-route [this params]
(str (secretary/render-route this) "?"
(secretary/encode-query-params params))))
(secretary/render-route (User. 1))
;; => "/users/1"
(secretary/render-route (User. 1) {:action :delete})
;; => "/users/1?action=delete"
IRouteMatches
It is seldom you will ever need to create your own route matching
implementation as the built in String
and RegExp
routes matchers
should be fine for most applications. Still, if you have a suitable
use case then this protocol is available. If your intention is to is
to use it with defroute
your implementation must return a map or
vector.
goog.History
(ns example
(:require [secretary.core :as secretary :refer-macros [defroute]]
[goog.events :as events])
(:import [goog History]
[goog.history EventType]))
(def application
(js/document.getElementById "application"))
(defn set-html! [el content]
(aset el "innerHTML" content))
(secretary/set-config! :prefix "#")
;; /#/
(defroute home-path "/" []
(set-html! application "<h1>OMG! YOU'RE HOME!</h1>"))
;; /#/users
(defroute users-path "/users" []
(set-html! application "<h1>USERS!</h1>"))
;; /#/users/:id
(defroute user-path "/users/:id" [id]
(let [message (str "<h1>HELLO USER <small>" id "</small>!</h1>")]
(set-html! application message)))
;; /#/777
(defroute jackpot-path "/777" []
(set-html! application "<h1>YOU HIT THE JACKPOT!</h1>"))
;; Catch all
(defroute "*" []
(set-html! application "<h1>LOL! YOU LOST!</h1>"))
;; Quick and dirty history configuration.
(doto (History.)
(events/listen EventType.NAVIGATE #(secretary/dispatch! (.-token %)))
(.setEnabled true))
Distributed under the Eclipse Public License, the same as Clojure.