# Write a function that reads the Simpsons phonebook from
# https://storage.googleapis.com/cs1111/practice/simpsons_phone_book.txt
# The function then finds and returns all phone numbers
# of people from Simpsons TV series.
# Assuming no area code included.
# (hint: use regular expression).

# For practice purpose, download the simpsons_phone_book.txt
# and write another version of the function that will
# read the local file.

import re
import urllib.request

def find_all_phone_number_from_url(url, regex):
    stream = urllib.request.urlopen(url).read().decode('UTF-8')

    phone_list = []

    matches = regex.finditer(stream)
    for match in matches:
        print(match)
        if match.group() not in phone_list:
            phone_list.append(match.group())

    return phone_list

url = "https://storage.googleapis.com/cs1111/practice/simpsons_phone_book.txt"
regex = re.compile(r"[0-9]?[0-9][0-9][0-9]-[0-9][0-9][0-9][0-9]")   # raw data, compile
print(find_all_phone_number_from_url(url, regex))



####################

def find_all_phone_number_from_file(filename, regex):
    infile = open(filename, "r")   # read mode

    phone_list = []

    for line in infile:
        obj = regex.search(line)     # look in the string (line), search for this pattern
                 # if found, search() returns the match object, otherwise None
        if obj != None:
            # print(obj, obj.group(), obj.start(), obj.end())
            if obj.group() not in phone_list:
                phone_list.append(obj.group())

    infile.close()
    return phone_list

regex = re.compile(r"[0-9]?[0-9][0-9][0-9]-[0-9][0-9][0-9][0-9]")   # raw data, compile
print(find_all_phone_number_from_file("simpsons_phone_book.txt", regex))