find.find()

Searches for text in a tab.

You can use this function to search normal HTTP(S) web pages. It searches a single tab: you can specify the ID of a particular tab to search, or it will search the active tab by default. It searches all frames in the tab.

You can make the search case-sensitive and make it match whole words only.

By default, the function just returns the number of matches found. By passing in the includeRangeData and includeRectData options, you can get more information about the location of the matches in the target tab.

This function stores the results internally, so the next time any extension calls find.highlightResults(), then the results of this find call will be highlighted, until the next time someone calls find().

This is an asynchronous function that returns a Promise.

Syntax

js
browser.find.find(
  queryPhrase,       // string
  options            // optional object
)

Parameters

options Optional

object. An object specifying additional options. It may take any of the following properties, all optional:

caseSensitive

boolean. If true the, the search is case-sensitive. Defaults to false.

entireWord

boolean. Match only entire words: so "Tok" will not be matched inside "Tokyo". Defaults to false.

includeRangeData

boolean. Include range data in the response, which describe where in the page DOM the match was found. Defaults to false.

includeRectData

boolean. Include rectangle data in the response, which describes where in the rendered page the match was found. Defaults to false

matchDiacritics

boolean. If true, the search distinguishes between accented letters and their base letters. For example, when set to true, searching for "résumé" does not find a match for "resume". Defaults to false.

tabId

integer. ID of the tab to search. Defaults to the active tab.

queryPhrase

string. The text to search for.

Return value

A Promise that will be fulfilled with an object containing up to three properties:

count

integer. The number of results found.

rangeData Optional

array. If includeRangeData was given in the options parameter, then this property will be included. It is provided as an array of RangeData objects, one for each match. Each RangeData object describes where in the DOM tree the match was found. This would enable, for example, an extension to get the text surrounding each match, so as to display context for the matches.

The items correspond to the items given in rectData, so rangeData[i] describes the same match as rectData[i].

Each RangeData contains the following properties:

endOffset

The ordinal position of the end of the match within its text node.

endTextNodePos

The ordinal position of the text node in which the match ended.

framePos

The index of the frame containing the match. 0 corresponds to the parent window. Note that the order of objects in the rangeData array will sequentially line up with the order of frame indexes: for example, framePos for the first sequence of rangeData objects will be 0, framePos for the next sequence will be 1, and so on.

startOffset

The ordinal position of the start of the match within its text node.

startTextNodePos

The ordinal position of the text node in which the match started.

rectData Optional

array. If includeRectData was given in the options parameter, then this property will be included. It is an array of RectData objects. It contains client rectangles for all the text matched in the search, relative to the top-left of the viewport. Extensions can use this to provide custom highlighting of the results.

Each RectData object contains rectangle data for a single match. It has two properties:

rectsAndTexts

An object containing two properties, both arrays:

  • rectList: an array of objects which each have four integer properties: top, left, bottom, right. These describe a rectangle relative to the top-left of the viewport.
  • textList: an array of strings, corresponding to the rectList array. The entry at textList[i] contains the part of the match bounded by the rectangle at rectList[i].

For example, consider part of a web page that looks like this:

Text reading "this domain is established to be used for illustrative examples in documents. You may use this domain in examples without prior coordination or asking for permission." and a "More information" link.

If you search for "You may", the match needs to be described by two rectangles:

This domain is established to be used for illustrative examples in documents. You may use this domain in examples without prior coordination or asking for permission.". The words "you may" are highlighted.

In this case, in the RectData that describes this match, rectsAndTexts.rectList and rectsAndTexts.textList will each have 2 items.

  • textList[0] will contain "You ", and rectList[0] will contain its bounding rectangle.
  • textList[1] will contain "may", and rectList[1] will contain its bounding rectangle.
text

The complete text of the match, "You may" in the example above.

Browser compatibility

BCD tables only load in the browser

Examples

Basic examples

Search the active tab for "banana", log the number of matches, and highlight them:

js
function found(results) {
  console.log(`There were: ${results.count} matches.`);
  if (results.count > 0) {
    browser.find.highlightResults();
  }
}

browser.find.find("banana").then(found);

Search for "banana" across all tabs (note that this requires the "tabs" permission or matching host permissions, because it accesses tab.url):

js
async function findInAllTabs(allTabs) {
  for (const tab of allTabs) {
    const results = await browser.find.find("banana", { tabId: tab.id });
    console.log(`In page "${tab.url}": ${results.count} matches.`);
  }
}

browser.tabs.query({}).then(findInAllTabs);

Using rangeData

In this example the extension uses rangeData to get the context in which the match was found. The context is the complete textContent of the node in which the match was found. If the match spanned nodes, the context is the concatenation of the textContent of all spanned nodes.

Note that for simplicity, this example doesn't handle pages that contain frames. To support this you'd need to split rangeData into groups, one per frame, and execute the script in each frame.

The background script:

js
// background.js

async function getContexts(matches) {
  // get the active tab ID
  const activeTabArray = await browser.tabs.query({
    active: true,
    currentWindow: true,
  });
  const tabId = activeTabArray[0].id;

  // execute the content script in the active tab
  await browser.tabs.executeScript(tabId, { file: "get-context.js" });
  // ask the content script to get the contexts for us
  const contexts = await browser.tabs.sendMessage(tabId, {
    ranges: matches.rangeData,
  });
  for (const context of contexts) {
    console.log(context);
  }
}

browser.browserAction.onClicked.addListener((tab) => {
  browser.find.find("example", { includeRangeData: true }).then(getContexts);
});

The content script:

js
/**
 * Get all the text nodes into a single array
 */
function getNodes() {
  const walker = document.createTreeWalker(
    document,
    window.NodeFilter.SHOW_TEXT,
    null,
    false,
  );
  const nodes = [];
  while ((node = walker.nextNode())) {
    nodes.push(node);
  }

  return nodes;
}

/**
 * Gets all text nodes in the document, then for each match, return the
 * complete text content of nodes that contained the match.
 * If a match spanned more than one node, concatenate the textContent
 * of each node.
 */
function getContexts(ranges) {
  const contexts = [];
  const nodes = getNodes();

  for (const range of ranges) {
    let context = nodes[range.startTextNodePos].textContent;
    let pos = range.startTextNodePos;
    while (pos < range.endTextNodePos) {
      pos++;
      context += nodes[pos].textContent;
    }
    contexts.push(context);
  }
  return contexts;
}

browser.runtime.onMessage.addListener((message, sender, sendResponse) => {
  sendResponse(getContexts(message.ranges));
});

Using rectData

In this example the extension uses rectData to "redact" the matches, by adding black DIVs over the top of their bounding rectangles:

Three search results with some texted redacted by black rectangles.

Note that in many ways this is a poor way to redact pages.

The background script:

js
// background.js

async function redact(matches) {
  // get the active tab ID
  const activeTabArray = await browser.tabs.query({
    active: true,
    currentWindow: true,
  });
  const tabId = activeTabArray[0].id;

  // execute the content script in the active tab
  await browser.tabs.executeScript(tabId, { file: "redact.js" });
  // ask the content script to redact matches for us
  await browser.tabs.sendMessage(tabId, { rects: matches.rectData });
}

browser.browserAction.onClicked.addListener((tab) => {
  browser.find.find("banana", { includeRectData: true }).then(redact);
});

The content script:

js
// redact.js

/**
 * Add a black DIV where the rect is.
 */
function redactRect(rect) {
  const redaction = document.createElement("div");
  redaction.style.backgroundColor = "black";
  redaction.style.position = "absolute";
  redaction.style.top = `${rect.top}px`;
  redaction.style.left = `${rect.left}px`;
  redaction.style.width = `${rect.right - rect.left}px`;
  redaction.style.height = `${rect.bottom - rect.top}px`;
  document.body.appendChild(redaction);
}

/**
 * Go through every rect, redacting them.
 */
function redactAll(rectData) {
  for (const match of rectData) {
    for (const rect of match.rectsAndTexts.rectList) {
      redactRect(rect);
    }
  }
}

browser.runtime.onMessage.addListener((message) => {
  redactAll(message.rects);
});

Example extensions