Putting it all together

We already extracted the entities and recognized the intent in several ways. We're now ready to put it all together to calculate a semantic representation for a user utterance!

1. We'll process the example dataset utterance:

   show me flights from denver to philadelphia on tuesday

   We'll hold a dictionary object to hold the result. The result will include the entities and the intent.

2. Let's extract the entities:

   import spacy
   from spacy.matcher import Matcher  
   nlp = spacy.load("en_core_web_md")
   matcher = Matcher(nlp.vocab)
   pattern = [{"POS": "ADP"}, {"ENT_TYPE": "GPE"}]
   matcher.add("prepositionLocation", [pattern])
   # Location entities
   doc = nlp("show me flights from denver to philadelphia on tuesday")
   matches = matcher(doc)
   for mid, start, end in matches:
       print(doc[start:end])
   ... 
   from denver
   to philadelphia
   # All entities:
   ents = doc...