How to replace strings in sentence with appropriate dictionary values?

Question

I have a dictionary as follows:

dict_ = { 
        'the school in USA' : 'some_text_1',
        'school' : 'some_text_1',
        'the holy church in brisbane' : 'some_text_2',
        'holy church' : 'some_text_2'
}

and a list of sentences as follows:

text_sent = ["Ram is going to the holy church in brisbane",\
             "John is going to holy church", \
             "shena is going to the school in USA", \
             "Jennifer is going to the school"]

I want to replace the occurrences of keys of dict_ dictionary with corresponding values in text_sent. I did this as follows:

for ind, text in enumerate(text_sent) :
    for iterator in dict_.keys() :
        if iterator in text : 
            text_sent[ind] = re.sub(iterator, dict_[iterator], text)

for i in text_sent:
    print(i)

Output I got is as follows:

Ram is going to the some_text_2 in brisbane
John is going to some_text_2
shena is going to the some_text_1 in USA
Jennifer is going to the some_text_1

Expected output is:

Ram is going to some_text_2
John is going to some_text_2
shena is going to some_text_1
Jennifer is going to some_text_1

What I need is, the strings that are longer (for example, " the holy church in brisbane ") need to be replaced, if in case, the complete string is not available in the sentence, only then the smaller version(for example, ' holy church ') should be used instead of the longer one for replacing corresponding value in text_sent's sentences.

Answer 1

You can use re.sub to make the replacements, using str.join to format the regex expression from the substring dictionary:

import re
d = {'the school in USA': 'some_text_1', 'school': 'some_text_1', 'the holy church in brisbane': 'some_text_2', 'holy church': 'some_text_2'}
text_sent = ["Ram is going to the holy church in brisbane",\
         "John is going to holy church", \
         "shena is going to the School in USA", \
         "Jennifer is going to the school"]

r = [re.sub('|'.join(d), lambda x:d[x.group()], i, re.I) for i in text_sent]

Output:

['Ram is going to some_text_2', 'John is going to some_text_2', 'shena is going to some_text_1', 'Jennifer is going to the some_text_1']

Answer 2

You can create an auxiliary list for the dict and sort it dependending on it's elements length.

dict_ = {'the school in USA' : 'some_text_1',
         'school' : 'some_text_1',
         'the holy church in brisbane' : 'some_text_2',
         'holy church' : 'some_text_2'}

text_sent = ["Ram is going to the holy church in brisbane",
             "John is going to holy church",
             "shena is going to the school in USA",
             "Jennifer is going to the school"]

dict_keys = list(dict_.keys())
dict_keys.sort(key=len)
dict_keys.reverse()

text_sent_replaced = []
for text in text_sent:
    modified_text = text
    for key in dict_:
        modified_text = modified_text.replace(key,dict_[key])
    text_sent_replaced.append(modified_text)

print(text_sent_replaced)

Answer 3

The main issue is that you didn't add a break statement. You are overriding values if there are multiple matches later on in the dict_ dictionary. Try this:

for ind, text in enumerate(text_sent) :
    for iterator in dict_.keys() :
        if iterator in text :
            text_sent[ind] = re.sub(iterator, dict_[iterator], text)
            break

Answer 4

This will accomplish the task without using re as long as the substituted elemets are at the end of each line, as was the case in your example:

for ind, text in enumerate(text_sent) :
    for iterator in dict_.keys() :
        if iterator in text :
            text_sent[ind] = text.split(iterator)[0] + dict_[iterator]

for i in text_sent:
    print(i)

#Prints:
#Ram is going to the some_text_2
#John is going to some_text_2
#shena is going to the some_text_1
#Jennifer is going to the some_text_1