Ability to extract content into json based on tag or document type

[dsabella84]

Is there a way to extract line items or keywords and store them into the contents field as json similar to parsio?

[stumpylog]

I suppose either a post processing script or via the API, but nothing is built in. The content is meant to be just that, the content of the document.

[dsabella84]

What are some thoughts on after the ocr process, run the content against a model to generate a json data. That data would be attributed to a document aside from the content field.

Process Flow:

1. Document pre processed
2. Document ran against specific model for data extraction
3. json data added to field on document
4. Document post processing

The idea is that I would like to have that json data for the post processing so i could send that to another service in my cluster.

[shamoon]

#2900 so I think this thread is essentially a duplicate.

I'll also just note this is not a trivial thing to do and in some ways would represent a paradigm shift for the software.

[dsabella84]

I agree this is not a trivial task, aside from #2900 I was thinking along of the lines that the model would still be local to the instance.

I see your point about a paradigm shift here but would be a nice to have feature for a local ecosystem of document management.

[github-actions[bot]]

This discussion has been automatically closed due to inactivity. Please see our contributing guidelines for more details.