Produce the most unique elements with least lists - python-3.x

Question

I am new here. I hope I can explain briefly below after giving an example.
example1: What is your name?
example1: Where are you from?
example1: How are you doing?
example2: What is your name?
example2: Where are you from?
example2: How are you doing?
example2: When did you move here?
example9: What is your name?
example3: Where are you from?
example23: Who gave you this book?
In the above example, I would like to print the unique questions by considering the number of example. So trying something like
expected output
example2: What is your name?
example2: Where are you from?
example2: How are you doing?
example2: When did you move here?
example23: Who gave you this book?
Here, I am searching for the unique questions in a file by considering fewer examples.
I played around something and placing that below.
import collections
s = collections.defaultdict(list)
u_s = set()
with open ('file.txt', 'r') as s1:
for line in s1:
data = line.split(':', maxsplit=1)
start = data[0]
end = data[-1]
if end not in u_s:
u_s.add(end)
s[start] += [end]
for start, ends in s.items():
print(start, ends[0])
for end in ends[1:]:
print(start, end)
Result that I am getting:
example1 What is your name?
example1 Where are you from?
example1 How are you doing?
example2 When did you move here?
example23 Who gave you this book?
Here, Instead of going to print example1, I want to consider example2 because it is giving more questions.
I tried by sorting the lines based on the repetitions of the line. I couldn't pass through it. I appreciate your help. Thanks

Answer 1

What your code achieved is to print all unique questions but cannot compare or print them in a whole set.
Apart from sorting, I would formulate the problem as to compare the combinations of example sets and select the one that contains the most unique questions with the least sets, so your question is more about the algorithm to me.
import collections
def calculate_contrib(values, set):
'''To calculate the contribution on the unique questions' number, based on values to add.
values: the list of question set to choose.
set: the already-added question set.'''
contrib = 0
for value in values:
if value not in set:
contrib += 1
return contrib
def print_result(x):
'''To print the result, x, as a dictionary, without repetition.'''
u_s = set()
for key, values in x.items():
for value in values:
if value not in u_s:
print(key,value)
u_s.add(value)
s = collections.defaultdict(list)
# get all questions in examples
with open('file.txt', 'r') as s1:
for line in s1:
data = line.split(':', maxsplit=1)
start = data[0]
end = data[-1]
s[start] += [end]
# Get the initial contribution on the unique questions' number for each example set
contrib = dict()
u_s = set()
result = dict()
for key,values in s.items():
contrib.update({key: calculate_contrib(s[key], u_s)})
# Execute the while loop when there are unique questions to add to u_s
while not(all([x == 0 for x in contrib.values()])):
# Add the example set with maximum contribution
max_contrib = 0
max_key = ""
for key, value in contrib.items():
if max_contrib < value:
max_key = key
max_contrib = value
result.update({max_key: s[max_key]})
u_s.update(s[max_key])
del s[max_key]
del contrib[max_key]
for key, values in s.items():
contrib[key] = calculate_contrib(values, u_s)
# print the result
print_result(result)
Above is a straightforward implementation, that is adding the example set with the most increase on the unique's number each time until no unique question remains.
Further improvement can be conducted. Hope it could give you some insight.