Question about NER and relation annotation

I am new to spacy and prodigy tools. I am able to perform domain-specific custom NER model but unable to extract relations/dependencies between entities using spacy via Jupyter notebook. The command I have used to export annotated data from prodigy to spacy is
% prodigy data-to-spacy ./outputfile name --lang en --ner name of annotated dataset name
is the above command is correct for custom NER and rel.manual annoations perfromed combined using
prodigy rel.manual ner_rels blank:en ./data.jsonl --label SUBJECT,LOCATION --span-label PERSON,GPE

If yes then please help me how can I use spacy to test my model for entities relations?

Thank you, kindly.

Zaid Kamil

Hi! spaCy doesn't have a built-in component for relation extraction at the moment, so there's no component you can train with it out-of-the-box and you'd have to use your own custom implementation if that's what you want to predict.

The following tutorial and example project shows how a relation extraction component can be implemented and what to consider for it – however, it's definitely a more advanced use case so if you're still new to spaCy, you might want to start with the more general usage docs and tutorials first :slightly_smiling_face:

Thanks, I will go through the suggested video.

Best regards
Zaid Kamil