PDF Parse Workflow with Docling and Elastic Search

[stangirala]

Context

Create a new workflow example that uses Docling for the pdf parsing and Elastic Search as the embedding query engine. This example is similar to what we have an offers an OSS alternative.

What

• Create a new workflow that uses Docling.
• Create steps to write generated embeddings to ES.
• Modify docker compose to start a local ES container to demonstrate the example.

Next Step

• More of an FYI, the image.py object and the requirements file need to be brought in sync once the image builder setup is working.

Testing

• Ran the workflow locally to test the write.
• Ran the workflow with indexify-server.
• Verified that the newly created ES indexes are query-able using embedding vector KNN.

Contribution Checklist

• If the python-sdk was changed, please run make fmt in python-sdk/.
• If the server was changed, please run make fmt in server/.
• Make sure all PR Checks are passing.