spaCy has a built-in tool called displacy for displaying simple, but clean and powerful visualizations. It offers two primary modes: named entity recognition and dependency parsing. Here, we will use the dep, or dependency mode:
displacy.render(doc, style='dep', jupyter=True)
Let's take the first sentence for a quick study: we can see that instrument is amod, or adjectively modified by Indian classicial. We pulled this phrase earlier as a noun chunk:
This means that when we pulled noun phrase chunks out of this sentence, spaCy must have finished dependency parsing already under the hood.
Also, notice the direction ...