I have a web page that asks the user for a paragraph of text, then performs some operation on it. To demo it to lazy users, I'd like to add an "I feel lucky" button that will grab some random text from Wikipedia and populate the inputs.
How can I use Javascript to fetch a sequence of text from a random Wikipedia article?
I found some examples of fetching and parsing articles using the Wikipedia API, but they tend to be server side. I'm looking for a solution that runs entirely from the client and doesn't get scuppered by same origin policy.
Note random gibberish is not sufficient; I need human-readable sentences that make sense.
I have a web page that asks the user for a paragraph of text, then performs some operation on it. To demo it to lazy users, I'd like to add an "I feel lucky" button that will grab some random text from Wikipedia and populate the inputs.
How can I use Javascript to fetch a sequence of text from a random Wikipedia article?
I found some examples of fetching and parsing articles using the Wikipedia API, but they tend to be server side. I'm looking for a solution that runs entirely from the client and doesn't get scuppered by same origin policy.
Note random gibberish is not sufficient; I need human-readable sentences that make sense.
Share Improve this question edited May 23, 2017 at 12:34 CommunityBot 11 silver badge asked Mar 8, 2013 at 12:08 rkagererrkagerer 4,2842 gold badges28 silver badges30 bronze badges1 Answer
Reset to default 12My answer builds on the technique suggested here.
The tricky part is formulating the correct query string:
http://en.wikipedia/w/api.php?action=query&generator=random&prop=extracts&exchars=500&format=json&callback=onWikipedia
generator=randomselects a random pageprop=extractsandexchars=500retrieves a 500 character extractformat=jsonreturns JSON-formatted datacallback=causes that data to be wrapped in a function call so it can be treated like any other<script>and injected into your page (see JSONP), thus bypassing cross-domain barriers.requestidcan optionally be added, with a new value each time, to avoid stale results from the browser cache (required in IE9)
The page served by the query is something that looks like this (I've added whitespace for readability):
onWikipedia(
{"query":
{"pages":
{"12362520":
{"pageid":12362520,
"ns":0,
"title":"Power Building",
"extract":"<p>The <b>Power Building<\/b> is a historic mercial building in
the downtown of Cincinnati, Ohio, United States. Built in 1903, it
was designed by Harry Hake. It was listed on the National Register
of Historic Places on March 5, 1999. One week later, a group of
buildings in the northeastern section of downtown was named a
historic district, the Cincinnati East Manufacturing and Warehouse
District; the Power Building is one of the district's contributing
properties.<\/p>\n<h2> Notes<\/h2>"
} } } }
)
Of course you'll get a different article each time.
Here's a full, working example which you can try out on JSBin.
<HTML><BODY>
<p><textarea id="textbox" style="width:350px; height:150px"></textarea></p>
<p><button type="button" id="button" onclick="startFetch(100, 500)">
Fetch random Wikipedia extract</button></p>
<script type="text/javascript">
var textbox = document.getElementById("textbox");
var button = document.getElementById("button");
var tempscript = null, minchars, maxchars, attempts;
function startFetch(minimumCharacters, maximumCharacters, isRetry) {
if (tempscript) return; // a fetch is already in progress
if (!isRetry) {
attempts = 0;
minchars = minimumCharacters; // save params in case retry needed
maxchars = maximumCharacters;
button.disabled = true;
button.style.cursor = "wait";
}
tempscript = document.createElement("script");
tempscript.type = "text/javascript";
tempscript.id = "tempscript";
tempscript.src = "http://en.wikipedia/w/api.php"
+ "?action=query&generator=random&prop=extracts"
+ "&exchars="+maxchars+"&format=json&callback=onFetchComplete&requestid="
+ Math.floor(Math.random()*999999).toString();
document.body.appendChild(tempscript);
// onFetchComplete invoked when finished
}
function onFetchComplete(data) {
document.body.removeChild(tempscript);
tempscript = null
var s = getFirstProp(data.query.pages).extract;
s = htmlDecode(stripTags(s));
if (s.length > minchars || attempts++ > 5) {
textbox.value = s;
button.disabled = false;
button.style.cursor = "auto";
} else {
startFetch(0, 0, true); // retry
}
}
function getFirstProp(obj) {
for (var i in obj) return obj[i];
}
// This next bit borrowed from Prototype / hacked together
// You may want to replace with something more robust
function stripTags(s) {
return s.replace(/<\w+(\s+("[^"]*"|'[^']*'|[^>])+)?>|<\/\w+>/gi, "");
}
function htmlDecode(input){
var e = document.createElement("div");
e.innerHTML = input;
return e.childNodes.length === 0 ? "" : e.childNodes[0].nodeValue;
}
</script>
</BODY></HTML>
One downside of generator=random is you often get talk pages or generated content that are not actual articles. If anyone can improve the query string to limit it to quality articles, that would be great!