Feat/knowledge extractor

[marmg]

Status	Type	⚠️ Core Change	Issue
Ready	Feature	No

Summary

Introducing a new component, the knowledge extractor.

The knowledge extractors will extract, at the same time, the spans and relations in a “triple” format (subject, relation, object).

The first knowledge extractor implemented is KnowGL:

• Rossiello et al. (AAAI 2023)
• Mihindukulasooriya et al. (ISWC 2022)
• Original Model

Usage:

import zshot
import spacy

from zshot import PipelineConfig, displacy
from zshot.knowledge_extractor import KnowGL
from zshot.utils.mappings import spans_to_wikipedia

nlp = spacy.blank("en")
nlp_config = PipelineConfig(
    knowledge_extractor=KnowGL()
)
nlp.add_pipe("zshot", config=nlp_config, last=True)

text = "The Italian Space Agency’s Light Italian CubeSat for Imaging of Asteroids, or LICIACube, will fly by Dimorphos to capture images and video of the impact plume as it sprays up off the asteroid and maybe even spy the crater it could leave behind."
text = "CH2O2 is a chemical compound similar to Acetamide used in International Business Machines Corporation (IBM)."
doc = nlp(text)
displacy.render(doc, style='rel')
print(spans_to_wikipedia(doc._.spans))