aniruddha27

def rule2_mod(text,index):
    
    doc = nlp(text)

    phrase = ''
    
    for token in doc:
        
        if token.i == index:
            

aniruddha27

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

    pat = []
    
    # iterate over tokens
    for token in doc:
        phrase = ''

aniruddha27

# create a df containing sentence and its output for rule 1
row_list = []

# df2 contains all the sentences from all the speeches
for i in range(len(df2)):
    
    sent = df2.loc[i,'Sent']
    year = df2.loc[i,'Year']
    output = rule1(sent)
    dict1 = {'Year':year,'Sent':sent,'Output':output}

aniruddha27

# create a df containing sentence and its output for rule 1
row_list = []

for i in range(len(df3)):
    
    sent = df3.loc[i,'Sent']
    year = df3.loc[i,'Year']
    output = rule1(sent)
    dict1 = {'Year':year,'Sent':sent,'Output':output}
    row_list.append(dict1)

aniruddha27

# function for rule 1: noun(subject), verb, noun(object)
def rule1(text):
    
    doc = nlp(text)
    
    sent = []
    
    for token in doc:
        
        # if the token is a verb

aniruddha27

text = df3.loc[9,'Sent'] 
print(text)
doc = nlp(text)
displacy.render(doc, style='dep',jupyter=True)

aniruddha27

# function to check output percentage for a rule
def output_per(df,out_col):
    
    result = 0
    
    for out in df[out_col]:
        if len(out)!=0:
            result+=1
    
    per = result/len(df)

aniruddha27

from random import randint
def rand_sent(df):
    
    index = randint(0, len(df))
    print('Index = ',index)
    doc = nlp(df.loc[index,'Sent'][1:])
    displacy.render(doc, style='dep',jupyter=True)
    
    return index

aniruddha27

row_list = []
# df2 contains all sentences from all speeches
for i in range(len(df2)):
    sent = df2.loc[i,'Sent']
    
    if (',' not in sent) and (len(sent.split()) <= 15):
        
        year = df2.loc[i,'Year']
        length = len(sent.split())
        

aniruddha27

# rule to extract initiative name
def sent_subtree(text):
    
    # pattern match for schemes or initiatives
    patterns = [r'\b(?i)'+'plan'+r'\b',
           r'\b(?i)'+'programme'+r'\b',
           r'\b(?i)'+'scheme'+r'\b',
           r'\b(?i)'+'campaign'+r'\b',
           r'\b(?i)'+'initiative'+r'\b',
           r'\b(?i)'+'conference'+r'\b',