Add new DoclingServe Integration

[maxdswain]

Summary and motivation

Docling is an extremely popular document information extraction library and in production uses cases, docling-serve can be used to host Docling in a more performant scalable fashion.

Detailed design

Follow the design of the already existing docling-haystack introducing a new DoclingServeConverter component accepting a list of str/pathlib.Path and producing a dict[str, list[Document]] but instead using the Docling serve endpoints following Docling serve's documentation. An run_async method would also be good to implement as the requests to Docling serve with be blocking.

I considered extending the existing Docling integration, but it has large dependencies (e.g., pytorch) which are unneeded when sending requests to Docling serve.

Checklist

If the request is accepted, ensure the following checklist is complete before closing this issue.

Tasks

• The code is documented with docstrings and was merged in the main branch
• Docs are published at https://docs.haystack.deepset.ai/
• There is a Github workflow running the tests for the integration nightly and at every PR
• A new label named like integration:<your integration name> has been added to the list of labels for this repository
• The labeler.yml file has been updated
• The package has been released on PyPI
• An integration tile with a usage example has been added to https://github.com/deepset-ai/haystack-integrations
• The integration has been listed in the Inventory section of this repo README
• The feature was announced through social media

[julian-risch]

Great idea @maxdswain ! Have you considered opening the issue in https://github.com/docling-project/docling-haystack ?
I was also wondering if there was a way to extend the existing component instead of introducing a new DoclingServeConverter component. However, having two separate components is probably better for users and maintainers. And the large dependencies that you mention are definitely a good reason for keeping things separate too. In that case the new component should be released as a spearate pypi package too.