XMLOutputParser does not parse CDATA

[Aman-14]

Checked other resources

• I added a very descriptive title to this issue.
• I searched the LangChain.js documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.
• I am sure that this is a bug in LangChain.js rather than my code.
• The bug is not resolved by updating to the latest stable version of LangChain (or the specific integration package).

Example Code

import { XMLOutputParser } from "./output_parsers/xml.js";

const xml = `<?xml version="1.0" encoding="UTF-8"?>
<userProfile>
  <userID>12345</userID>
  <email>[email protected]</email>
  <bio><![CDATA[John is a senior developer with <10 years> of experience. He uses <Typescript> at work.]]></bio>
</userProfile>`;

const parser = new XMLOutputParser();
parser.parse(xml).then(console.log).catch(console.error);

Output:

{
  userProfile: [
    { userID: '12345' },
    { email: '[email protected]' },
    { bio: '' }
  ]
}

Expected Output:

{
  userProfile: [
    { userID: '12345' },
    { email: '[email protected]' },
    {
      bio: 'John is a senior developer with <10 years> of experience. He uses <Typescript> at work.'
    }
  ]
}

As we can see XMLOutputParser does not parse CDATA text.

Error Message and Stack Trace (if applicable)

No response

Description

The XMLOutputParser in the LangChain library is not correctly parsing XML content that includes CDATA sections when used in a Node.js environment. The parser appears to be ignoring the CDATA content, resulting in incomplete parsing of XML.

System Info

OS: mac os
Node version: 20.14.0
Yarn version: 1.22.22

[dosubot]

The issue with the XMLOutputParser not parsing CDATA sections correctly is due to the sax.parser used in the parseXMLMarkdown function not handling CDATA sections. The current implementation lacks an event handler for CDATA sections.

To fix this, you need to add an event handler for oncdata in the sax.parser configuration. Here is how you can modify the code:

const parser = sax.parser(true);

// Event handlers
parser.onopentag = (node: any) => { /* ... */ };
parser.onclosetag = () => { /* ... */ };
parser.ontext = (text: any) => { /* ... */ };
parser.onattribute = (attr: any) => { /* ... */ };

// Add handler for CDATA sections
parser.oncdata = (cdata: any) => {
  if (elementStack.length > 0) {
    const currentElement = elementStack[elementStack.length - 1];
    currentElement.text += cdata;
  }
};

This addition will allow the parser to correctly handle and include CDATA sections in the parsed result ^[1][2].

To continue talking to Dosu, mention @dosu.

[Aman-14]

Correct, i can create a PR to fix this. Please ping me if any human is reading this.