dom-highlight-lib

Highlight tokens on the DOM like the Grammarly. Inspired by Making Grammarly Feel Native On Every Website.

Install

yarn add dom-highlight-lib
# or
npm install --save dom-highlight-lib

Usage

1. Define an asynchronous match function that accepts an array of strings and returns an two-level array of tokens (the keywords found in the input strings).

   • For each input string, the function should return an array of tokens.
   • The order of returned two-level array should be same to the order of the input strings, so that we could tell what tokens are returned for a specific input string.

   // for example
   let id = 0;
   const regex = /\w+/mg
   const match = (texts) => new Promise((resolve) => {
     const tokenize = (text) => {
       const tokens = [];
       let m = regex.exec(text);
       while (m !== null) {
         tokens.push({
           id: `${id}`,
           color: `#ff0000`,
           start: m.index,
           end: m.index + m[0].length,
           keyword: m[0]
         });
         id += 1;
         m = regex.exec(text);
       }
       return tokens;
     };
   
     resolve(texts.map(tokenize));
   });
   

2. Decide what to do when the user mouse hover on a highlight.

   // for example
   const tokenView = document.createElement('div');
   tokenView.style.color = '#fff';
   tokenView.style.borderRadius = '8px';
   tokenView.style.boxShadow = '0 2px 15px 0 rgba(68, 70, 73, 0.2)';
   tokenView.style.padding = '10px';
   tokenView.style.position = 'absolute';
   tokenView.style.width = '362px';
   tokenView.style.zIndex = '2147483647';
   document.body.appendChild(tokenView);
   
   const showToken = (token, rect, event) => {
     tokenView.style.background = token.color;
     tokenView.style.top = `${rect.top}px`;
     tokenView.style.left = `${rect.left}px`;
     tokenView.innerHTML = `<ul><li>ID: {token.id}</li><li>Title: {token.keyword}</li></ul>`;
     tokenView.style.display = 'block';
   };
   
   const hideToken = (token, rect, event) => {
     tokenView.innerHTML = '';
     tokenView.style.display = 'none';
   };
   

3. Optionally, provide below parameters:

   • selectors: specify CSS selector string of the root nodes to match. Defaults to the document.documentElement.
   • ignoreInvisibleNodes: whether ignores invisible nodes. Defaults to true.
   • minTextLength: Ignore text nodes whose text length is less than specified value. Defaults to 3.
   • trimInvisibleChars: Whether ignores the leading and trailing invisible chars before counting the text length. Defaults to true.
   • minBatchTextLength: controls how often to invoke the match function according to the total number of characters in the strings. Defaults to 1000.
   • className: the CSS class name for the highlights container. Defaults to empty string.
   • throttleUpdates: Throttle highlight updates, update at most once every specified milliseconds. Defaults to 500ms.

4. Start the highlighter.

   // for example, on a web page
   const highlighter = require('dom-highlight-lib');
   highlighter.start({
     match,
     showToken,
     hideToken,
     selectors: '',
     ignoreInvisibleNodes: true,
     minTextLength: 3,
     trimInvisibleChars: true,
     minBatchTextLength: 1000,
     className: '',
     throttleUpdates: 500
   });
   

   The highlighter now scans and monitor changes (scroll, resize, and MutationObserver) on the web page, and update highlights when necessary.

TODO

• textarea and input controls are not supported because Range.getClientRects() returns empty list.
• Because the match function is asynchronous, it possible for the highlighter to run into an inconsistent state when DOM changes frequently.