nlp_ie_21.py

# function for rule 2
def rule2(text):
    
    doc = nlp(text)

    pat = []
    
    # iterate over tokens
    for token in doc:
        phrase = ''
        # if the word is a subject noun or an object noun
        if (token.pos_ == 'NOUN')\
            and (token.dep_ in ['dobj','pobj','nsubj','nsubjpass']):
            
            # iterate over the children nodes
            for subtoken in token.children:
                # if word is an adjective or has a compound dependency
                if (subtoken.pos_ == 'ADJ') or (subtoken.dep_ == 'compound'):
                    phrase += subtoken.text + ' '
                    
            if len(phrase)!=0:
                phrase += token.text
             
        if  len(phrase)!=0:
            pat.append(phrase)
        
    
    return pat