Chrome Extension: how to capture selected text and send to a web service

For the Google Chrome extension, I need to capture selected text in a web page and send to a web service. I'm stuck!

First I tried a bookmarklet, but Chrome on Mac seems to have some bookmarklet bugs so I decided to write an extension.

I use this code in my ext:

function getSelText(){
    var txt = 'nothing';
    if (window.getSelection){
 开发者_如何学运维       txt = "1" + window.getSelection();
    } else if (document.getSelection) {
        txt = "2" + document.getSelection();
    } else if (document.selection) {
        txt = "3" + document.selection.createRange().text;
    } else txt = "wtf";
    return txt;
}
var selection = getSelText();
alert("selection = " + selection);

When I click on my extension icon, I get a "1". So I think the act of selecting outside the browser window is causing the text to not be seen by the browser as "selected" any more.

Just a theory....

thoughts?

---

You can do this by using Extensions Messaging. Basically, your "background page" will send the request to your service. For example, lets say you have a "popup" and once you click on it, it will do a "Google search" which is your service.

content_script.js

In your content script, we need to listen for a request coming from your extension, so that we send it the selected text:

chrome.extension.onRequest.addListener(function(request, sender, sendResponse) {
    if (request.method == "getSelection")
      sendResponse({data: window.getSelection().toString()});
    else
      sendResponse({}); // snub them.
});

background.html

Now in background page you can handle the popup onclick event so that we know we clicked on the popup. Once we clicked on it, the callback fires, and then we can send a request to the content script using "Messaging" to fetch the selected text.

chrome.browserAction.onClicked.addListener(function(tab) {
  chrome.tabs.sendRequest(tab.id, {method: "getSelection"}, function(response){
     sendServiceRequest(response.data);
  });
});

function sendServiceRequest(selectedText) {
  var serviceCall = 'http://www.google.com/search?q=' + selectedText;
  chrome.tabs.create({url: serviceCall});
}

As you have seen, I registered a listener in a content script to allow my extension to send and receive messages from it. Then once I received a message, I handle it by searching for Google.

Hopefully, you can use what I explained above and apply it to your scenario. I just have to warn you that the code written above is not tested, so their might be spelling, or syntax errors. But those can easily be found by looking at your Inspector :)

---

content script

document.addEventListener('mouseup',function(event)
{
    var sel = window.getSelection().toString();

    if(sel.length)
        chrome.extension.sendRequest({'message':'setText','data': sel},function(response){})
})

Background Page

<script>
var seltext = null;

chrome.extension.onRequest.addListener(function(request, sender, sendResponse)
{
    switch(request.message)
    {
        case 'setText':
            window.seltext = request.data
        break;

        default:
            sendResponse({data: 'Invalid arguments'});
        break;
    }
});


function savetext(info,tab)
{
    var jax = new XMLHttpRequest();
    jax.open("POST","http://localhost/text/");
    jax.setRequestHeader("Content-Type","application/x-www-form-urlencoded");
    jax.send("text="+seltext);
    jax.onreadystatechange = function() { if(jax.readyState==4) { alert(jax.responseText);  }}
}

var contexts = ["selection"];
for (var i = 0; i < contexts.length; i++)
{
    var context = contexts[i];
    chrome.contextMenus.create({"title": "Send to Server", "contexts":[context], "onclick": savetext});  
}

</script>

manifest.json

{
  "name": "Word Reminder",
  "version": "1.0",
  "description": "Word Reminder.",
  "browser_action": {
    "default_icon": "images/stick-man1.gif",
    "popup":"popup.html"
  },

  "background_page": "background.html",

  "content_scripts": [
    {
        "matches": ["<all_urls>"],
      "js": ["js/myscript.js"]
    }
  ],

  "permissions": [
    "http://*/*",
    "https://*/*",
    "contextMenus",
    "tabs"
  ]
}

and here is the link where i have all in one extension to download. after reading this i tried of my own and have published.

and here is the complete source

http://vikku.info/programming/chrome-extension/get-selected-text-send-to-web-server-in-chrome-extension-communicate-between-content-script-and-background-page.htm

Enjoy

---

Using a content_scripts is not a great solution as it injection to all documents including iframe-ads etc. I get an empty text selection from other pages than the one I expect half the times on messy web sites.

A better solution is to inject code into the selected tab only, as this is where your selected text lives anyhow. Example of jquery doc ready section:

$(document).ready(function() {
    // set up an event listener that triggers when chrome.extension.sendRequest is fired.
    chrome.extension.onRequest.addListener(
        function(request, sender, sendResponse) {
            // text selection is stored in request.selection
            $('#text').val( request.selection );
    });

    // inject javascript into DOM of selected window and tab.
    // injected code send a message (with selected text) back to the plugin using chrome.extension.sendRequest
    chrome.tabs.executeScript(null, {code: "chrome.extension.sendRequest({selection: window.getSelection().toString() });"});
});

---

It is not clear from your code where it is. What I mean, is that if this code is either in popup html or background html then the results you are seeing are correct, nothing in those windows will be selected.

You will need to place this code in a content script so that it has access to the DOM of the page, and then when you click your browser action, you will need to send a message to the content script to fetch the current document selection.

---

You don't need a Google API for something as simple as this...

I'll use the Bing online service as an example. Note that the URL is set up to accept a parameter:

var WebService='http://www.bing.com/translator/?text='; 


frameID.contentWindow.document.body.addEventListener('contextmenu',function(e){

 T=frameID.contentWindow.getSelection().toString();

 if(T!==''){e.preventDefault(); Open_New_Tab(WebService+encodeURIComponent(T)); return false;} 

},false);

NB: The function "Open_New_Tab()" used above is an imaginary one that accepts the webservice URL with the encoded selected text as a parameter.

That's the idea basically.