cleaning.py

import re
import copy
import json

def make_json(rawtext: list):
    qbank = []
    q_template = {"question": None, "answers": []}
    current_q = copy.deepcopy(q_template) # need deep copy of template

    while len(rawtext) > 0:
        line = rawtext.pop(0)
        check_question = re.match(pattern, line)
        if check_question:
            if len(current_q['answers']) > 0:
                qbank.append(current_q)
                current_q = copy.deepcopy(q_template) 
            current_q["question"] = check_question.groups()[1].strip().lower()
        elif '▪' in line:
            current_q["answers"].append(re.sub('^[\s▪]+', '', line).strip().lower())
    qbank.append(current_q)
    
    return  {"Questions": qbank}

def json_to_lower(original: dict)-> dict:
    '''
    Used to lower original json, to avoid loosing extra added info.
    besides, lovely nested comprehension.
    '''
    
    questions = original['Questions']
    lquestions = [{'question':q['question'].lower(),
                   'answers':[x.lower() for x in q['answers']]} for q in questions]
    return {'Questions': lquestions}

with open('question_bank.txt') as raw:
    rawtext = raw.readlines()
rawtext = [line.replace('\t', '').strip() for line in rawtext]
rawtext = [line for line in rawtext if len(line) > 0]

cleaner = []
pattern = '^(\d+\.)(.+)$' # captures whatever comes after the question number dot pattern

while len(rawtext) >0 :
    line = rawtext.pop(0)
    if line[0].isdigit() and not re.match(pattern, line):
        cleaner.append(line + rawtext.pop(0))
    else:
        cleaner.append(line)

processed = make_json(cleaner)

with open('question_bank.json','w') as output:
    json.dump(processed, output, indent = 2)