Using Javascript regex for extracting JSON string content including property values within HTML code

Question

Using Javascript regex for extracting JSON string content including property values within HTML code

I have data in the following format that is not valid JSON. I am trying to convert it using regex, but the parse won't go through. For more details, refer to this question which I've tried.

The problem lies in the fact that the response string also contains HTML, variables, and trailing commas. Below is the test data that is failing:

`var _nr_metadata = {
    site_base_url : "http://newsrack.in",
    issue_name    : "iihs_feeds",
    category_name : "Chikungunya",
    listing_url   : "/stories/servelots/iihs_feeds/16"
  }
  var _nr_stories = [
    {
      title  : "Alarm bells ringing:194 dengue cases in 2 weeks in district",
      url    : "http://www.tribuneindia.com/news/ludhiana/alarm-bells-ringing-194-dengue-cases-in-2-weeks-in-district/486718.html",
      source : "The Tribune",
      date   : "25.10.2017",
      desc   : "Tribune News Service\nLudhiana, October 24\nThe number of dengue cases is rapidly increasing in the district as 194 confirmed cases have been recorded by the Health Department ..."
    },
    ...
]

The expected output should be a valid JSON for use in the rest of my application. Unfortunately, I do not have control over modifying the server response.

javascript json regex parsing jsonlint

Answer 1

Answer №1

This server doesn't actually function as a JSON API; rather, it produces JavaScript code. Even the slightest alteration to the response format, like using semi-colons to terminate statements, could disrupt an attempt to parse it successfully. Without knowledge of your specific use case, this approach seems risky.

Despite these challenges, attempting to parse it using JSON.parse appears to be less hazardous than other methods. I managed to get it to parse by implementing these replacements:

var text = document.getElementById('response').innerText;

// 1. Convert vars to properties
text = text.replace(/^\s*var\s+(\w+)\s*=/gm, ',"$1":');

// 2. Quote the keys
text = text.replace(/^\s*(\w+)\s*:/gm, '"$1":');

// 3. Get rid of the '' at the end
text = text.replace(/^\s*\},[\n\s]*''.*$/gm, '}');

// 4. Remove the comma that step 1 added at the start
text = text.replace(/^\s*,/g, '');

// 5. Wrap everything in an object
text = '{' + text + '}';

var obj = JSON.parse(text);

console.log(obj);

<script id="response" type="custom">
var _nr_metadata = {
    site_base_url : "http://newsrack.in",
    ...
    // Truncated for brevity
    ...
}
</script>

I have employed the m flag in multiple RegEx patterns assuming certain data starts at the beginning of a line, which might be too fragile depending on variations in your source data.

Answer 2

This server doesn't actually function as a JSON API; rather, it produces JavaScript code. Even the slightest alteration to the response format, like using semi-colons to terminate statements, could disrupt an attempt to parse it successfully. Without knowledge of your specific use case, this approach seems risky.

Despite these challenges, attempting to parse it using JSON.parse appears to be less hazardous than other methods. I managed to get it to parse by implementing these replacements:

var text = document.getElementById('response').innerText;

// 1. Convert vars to properties
text = text.replace(/^\s*var\s+(\w+)\s*=/gm, ',"$1":');

// 2. Quote the keys
text = text.replace(/^\s*(\w+)\s*:/gm, '"$1":');

// 3. Get rid of the '' at the end
text = text.replace(/^\s*\},[\n\s]*''.*$/gm, '}');

// 4. Remove the comma that step 1 added at the start
text = text.replace(/^\s*,/g, '');

// 5. Wrap everything in an object
text = '{' + text + '}';

var obj = JSON.parse(text);

console.log(obj);

<script id="response" type="custom">
var _nr_metadata = {
    site_base_url : "http://newsrack.in",
    ...
    // Truncated for brevity
    ...
}
</script>

I have employed the m flag in multiple RegEx patterns assuming certain data starts at the beginning of a line, which might be too fragile depending on variations in your source data.

Using Javascript regex for extracting JSON string content including property values within HTML code

Answer №1

Similar questions

What is the best way to customize the spacing of grid lines in chartist.js?

Transferring a JSON document to an Express server with JavaScript

What causes the delay in CSS animations?

What is the ideal framerate for smooth animation?

Error: You cannot implement an import statement beyond a module while utilizing reactjs CDN Links

How can I navigate through all the divs by setting an interval slider with a random starting point?

Navigating a Two-Dimensional Array in JavaScript

What is the best way to create a regular expression that can detect a string containing a full stop ('.') in the middle and ending with a comma?

Unable to alter the height of the element

Node js can sometimes be unreliable

Premature audio cutoffs in Javascript/CSS when using onmousedown event

An Iframe lacks the ability to showcase HTML content, unlike a browser which is capable of doing

Create a feature that allows users to view photos similar to the way it

Using Material UI's ProgressBar to create a countdown feature in a React

Concentrate on elements that are activated without the need for clicking

Is there a way to construct a table within MySql?

Using mappers to create a computed property for two-way binding

Struggling with integrating HTML Canvas within Vue.js

I'm new to React and curious - why do we need to use ReactDOM.render() to update HTML instead of just writing the HTML directly?

Instructions for successfully populating Excel without encountering error warnings by utilizing JSONArray and POI