TextractDocument.html() returns LAYOUT error even after Job has completed with Layout Info

Question

Periodically for some documents (PDFs) but I'm not able to understand why. Textract will not be able to determine the layout even though the layout information is in the blocks.

For example: 
```
new TextractDocument({ Blocks: blocks } as any).html();
```

Will Return
```
Error: Page.html() is not yet implemented for results where Textract LAYOUT was not enabled
```

**Inputs**

```
const cmd = new StartDocumentAnalysisCommand({
      DocumentLocation: {
        S3Object: {
          Bucket: bucket,
          Name: key,
        },
      },
      FeatureTypes: ['LAYOUT'],
    });
    
    const resy = await textract.send(command);
```

Alex T · Answer

Hi and thanks for raising this,

Since the `hasLayout` check looks for whether *each page* has any layout items, I think the most likely cause for this may be documents with *blank pages*?

Because this issue seems to be within the Open Source [Amazon Textract Response Parser for JavaScript/TypeScript](https://www.npmjs.com/package/amazon-textract-response-parser), it'd probably be helpful to raise it [on the project GitHub](https://github.com/aws-samples/amazon-textract-response-parser/issues). If you're able to confirm whether it's failing on only blank pages, or there are any other cases, that'd be super helpful information for the team to debug!

**Update, 2024-11-19:**

We've gone ahead and released a new version 0.4.3 of TRP.js to tackle the blank page issue - see the [PR on GitHub](https://github.com/aws-samples/amazon-textract-response-parser/pull/186) and the [version on NPM](https://www.npmjs.com/package/amazon-textract-response-parser/v/0.4.3). Please try upgrading when you can, and let us know if it solves the issue or there's something else going on!

TextractDocument.html() returns LAYOUT error even after Job has completed with Layout Info

Add your answer

Relevant content