nlp_ie_11.py

# to extract initiatives using pattern matching
def all_schemes(text,check):
    
    schemes = []
    
    doc = nlp(text)
    
    # initiatives
    prog_list = ['programme','scheme',
                 'initiative','campaign',
                 'agreement','conference',
                 'alliance','plan']
    
    # pattern to match initiatives names 
    pattern = [{'POS':'DET'},
               {'POS':'PROPN','DEP':'compound'},
               {'POS':'PROPN','DEP':'compound'},
               {'POS':'PROPN','OP':'?'},
               {'POS':'PROPN','OP':'?'},
               {'POS':'PROPN','OP':'?'},
               {'LOWER':{'IN':prog_list},'OP':'+'}
              ]
    
    if check == 0:
        # return blank list
        return schemes

    # Matcher class object 
    matcher = Matcher(nlp.vocab) 
    matcher.add("matching", None, pattern) 
    matches = matcher(doc)

    for i in range(0,len(matches)):
        
        # match: id, start, end
        start, end = matches[i][1], matches[i][2]
        
        if doc[start].pos_=='DET':
            start = start+1
        
        # matched string
        span = str(doc[start:end])
        
        if (len(schemes)!=0) and (schemes[-1] in span):
            schemes[-1] = span
        else:
            schemes.append(span)
        
    return schemes

# apply function
df2['Schemes1'] = df2.apply(lambda x:all_schemes(x.Sent,x.Check_Schemes),axis=1)