@studysync/html-to-react

html-to-react

A lightweight library that converts raw HTML to a React DOM structure.

Why?

I had a scenario where an HTML template was generated by a different team, yet I wanted to leverage React for the parts I did have control over. The template basically contains something like:

<div class="row">
  <div class="col-sm-6">
    <div data-report-id="report-1">
      <!-- A React component for report-1 -->
    </div>
  </div>
  <div class="col-sm-6">
    <div data-report-id="report-2">
      <!-- A React component for report-2 -->
    </div>
  </div>
</div>

I had to replace each <div> that contains a data-report-id attribute with an actual report, which was nothing more than a React component.

Simply replacing the <div> elements with a React component would end up with multiple top-level React components that have no common parent.

The html-to-react module solves this problem by parsing each DOM element and converting it to a React tree with one single parent.

Installation

$ npm install --save html-to-react

Examples

Simple

The following example parses each node and its attributes and returns a tree of React elements.

const ReactDOMServer = require('react-dom/server');
const HtmlToReactParser = require('html-to-react').Parser;

const htmlInput = '<div><h1>Title</h1><p>A paragraph</p></div>';
const htmlToReactParser = new HtmlToReactParser();
const reactElement = htmlToReactParser.parse(htmlInput);
const reactHtml = ReactDOMServer.renderToStaticMarkup(reactElement);

assert.equal(reactHtml, htmlInput); // true

With Custom Processing Instructions

If certain DOM nodes require specific processing, for example if you want to capitalize each <h1> tag, the following example demonstrates this:

const ReactDOMServer = require('react-dom/server');
const HtmlToReact = require('html-to-react');
const HtmlToReactParser = require('html-to-react').Parser;

const htmlInput = '<div><h1>Title</h1><p>Paragraph</p><h1>Another title</h1></div>';
const htmlExpected = '<div><h1>TITLE</h1><p>Paragraph</p><h1>ANOTHER TITLE</h1></div>';

const isValidNode = function () {
  return true;
};

// Order matters. Instructions are processed in the order they're defined
const processNodeDefinitions = new HtmlToReact.ProcessNodeDefinitions(React);
const processingInstructions = [
  {
    // Custom <h1> processing
    shouldProcessNode: function (node) {
      return node.parent && node.parent.name && node.parent.name === 'h1';
    },
    processNode: function (node, children) {
      return node.data.toUpperCase();
    }
  },
  {
    // Anything else
    shouldProcessNode: function (node) {
      return true;
    },
    processNode: processNodeDefinitions.processDefaultNode
  }
];
const htmlToReactParser = new HtmlToReactParser();
const reactComponent = htmlToReactParser.parseWithInstructions(htmlInput, isValidNode,
  processingInstructions);
const reactHtml = ReactDOMServer.renderToStaticMarkup(reactComponent);
assert.equal(reactHtml, htmlExpected);

Replace the Children of an Element

There may be a situation where you want to replace the children of an element with a React component. This is beneficial if you want to:

• a) Preserve the containing element
• b) Not rely on any child node to insert your React component

Example

Below is a simple template that could get loaded via ajax into your application

Before

<div class="row">
  <div class="col-sm-6">
    <div data-container="wysiwyg">
      <h1>Sample Heading</h1>
      <p>Sample Text</p>
    </div>
  </div>
</div>

After

You may want to extract the inner html from the data-container attribute, store it and then pass it as a prop to your injected RichTextEditor.

<div class="row">
  <div class="col-sm-6">
    <div data-container="wysiwyg">
      <RichTextEditor html={"<h1>Sample heading</h1><p>Sample Text</p>"} />
    </div>
  </div>
</div>

Setup

In your instructions object, you must specify replaceChildren: true.

const React = require('react');
const HtmlToReact = require('html-to-react');
const HtmlToReactParser = require('html-to-react').Parser;

const htmlToReactParser = new HtmlToReactParser();
const htmlInput = '<div><div data-test="foo"><p>Text</p><p>Text</p></div></div>';
const htmlExpected = '<div><div data-test="foo"><h1>Heading</h1></div></div>';

const isValidNode = function () {
  return true;
};

const processNodeDefinitions = new HtmlToReact.ProcessNodeDefinitions(React);

// Order matters. Instructions are processed in
// the order they're defined
const processingInstructions = [
  {
    // This is REQUIRED, it tells the parser
    // that we want to insert our React
    // component as a child
    replaceChildren: true,
    shouldProcessNode: function (node) {
      return node.attribs && node.attribs['data-test'] === 'foo';
    },
    processNode: function (node, children, index) {
      return React.createElement('h1', {key: index,}, 'Heading');
    }
  },
  {
    // Anything else
    shouldProcessNode: function (node) {
      return true;
    },
    processNode: processNodeDefinitions.processDefaultNode,
  },
];

const reactComponent = htmlToReactParser.parseWithInstructions(
  htmlInput, isValidNode, processingInstructions);
const reactHtml = ReactDOMServer.renderToStaticMarkup(
  reactComponent);
assert.equal(reactHtml, htmlExpected);

With Preprocessing Instructions

There may be situations where you want to preprocess nodes before rendering them, analogously to the custom processing instructions functionality. The rationale for supporting preprocessing hooks is generally that you might want to apply more general processing to nodes, before applying custom processing hooks to filtered sets of nodes. You could accomplish the same by calling common functions from your custom processing hooks, but the preprocessing hooks API makes it more convenient.

Example

Below is a simple template in which you may want to replace div IDs, via a preprocessing hook:

<div class="row">
  <div id="first" data-process="shared">
    <p>Sample For First</p>
  </div>
  <div id="second" data-process="shared">
    <p>Sample For Second</p>
  </div>
</div>

We want to process the above HTML into the following:

<div class="row">
  <h1 id="preprocessed-first">First</h1>
  <h2 id="preprocessed-second">Second</h2>
</div>

We can accomplish that with the following script, using a combination of preprocessing and custom processing instructions:

const React = require('react');
const HtmlToReact = require('html-to-react');
const HtmlToReactParser = require('html-to-react').Parser;

const htmlToReactParser = new HtmlToReactParser();
const htmlInput = '<div class="row">' +
  '<div id="first" data-process="shared">' +
    '<p>Sample For First</p>' +
  '</div>' +
  '<div id="second" data-process="shared">' +
    '<p>Sample For Second</p>' +
  '</div>' +
'</div>';

const htmlExpected = '<div class="row">' +
  '<h1 id="preprocessed-first">First</h1>' +
  '<h2 id="preprocessed-second">Second</h2>' +
'</div>';

const isValidNode = function () {
  return true;
};

const preprocessingInstructions = [
  {
    shouldPreprocessNode: function (node) {
      return node.attribs && node.attribs['data-process'] === 'shared';
    },
    preprocessNode: function (node) {
      node.attribs = {id: `preprocessed-${node.attribs.id}`,};
    },
  }
];
const processNodeDefinitions = new HtmlToReact.ProcessNodeDefinitions(React);
const processingInstructions = [
  {
    shouldProcessNode: function (node) {
      return node.attribs && node.attribs.id === 'preprocessed-first';
    },
    processNode: function(node, children, index) {
      return React.createElement('h1', {key: index, id: node.attribs.id,}, 'First');
    },
  },
  {
    shouldProcessNode: function (node) {
      return node.attribs && node.attribs.id === 'preprocessed-second';
    },
    processNode: function (node, children, index) {
      return React.createElement('h2', {key: index, id: node.attribs.id,}, 'Second');
    },
  },
  {
    shouldProcessNode: function (node) {
      return true;
    },
    processNode: processNodeDefinitions.processDefaultNode,
  },
];

const reactComponent = parser.parseWithInstructions(htmlInput, isValidNode, processingInstructions,
  preprocessingInstructions);
const reactHtml = ReactDOMServer.renderToStaticMarkup(reactComponent);
assert.equal(reactHtml, htmlExpected);

Tests & Coverage

Test locally: $ npm test

Test with coverage and report coverage to Coveralls: $ npm run test-coverage

Test with coverage and open HTML report: $ npm run test-html-coverage