Option to not overwrite existing kb_id in entity_linker

[kinghuang]

The EntityLinker sets the kb_id for all entities, regardless of whether the entity has a kb_id or not. This is inconvenient in pipelines where some kb_ids may already be set by an earlier pipeline component.

I'd like an option in the entity_linker to not overwrite existing kb_ids. This could be like the entity_recognizer, which never overwrites existing entity labels, or entity_ruler, which has a config about overwriting entity labels.

[polm]

Thanks for the suggestion. We've recently added overwrite-related settings to the Morphologizer, and as part of that discussion we realized it would be good to have that option uniformly for all components, so we'd definitely be up for adding this.

[adrianeboyd]

You're in luck: this has already been implemented in #9050, it will be in v3.2.0.

[kinghuang]

Awesome. Thanks, all! 🎉