import csv
import numpy as np

def get_data(filename):
    # This functions returns the data as a list of rows
    # where each row has the form [1, Belem, oil] with
    # a transaction number, location, and product
    with open(filename,'r') as data_file:
        csv_reader = csv.reader(data_file, delimiter=',')
        data = []
        for row in csv_reader:
            trans_num = int(row[0].strip())
            city = row[1]
            item = row[2]
            data.append([trans_num, city, item])
        return data

def get_cities(data):
    # We'll exploit the set() feature in Python to ensure
    # that duplicates are first removed before a list is made.
    # Start with a set:
    cities_set = set()
    for row in data:
        # Add to the set, which automatically removes duplicates
        cities_set.add(row[1])
    # Convert to list:
    cities = list(cities_set)
    return cities

def get_products(data):
    # Write your code here to return a list of all products.
    # First use a set and then convert to list.

    return product_list

def sort_products(products):
    sorted_products = sorted(products, key=lambda x:x[0])
    return sorted_products

def get_purchases(data):
    # We've written this so to return a list of transactions
    # where every transaction has the form
    # [1, ['oil', 'cheese', 'banana', 'egg', 'chicken', 'garlic']]
    # Thus, the first element is the transaction number and
    # the second is itself a list (of products)
    transactions = []
    current_trans = data[0][0]
    current_set = []
    for row in data:
        if row[0] != current_trans:
            transactions.append([current_trans, current_set])
            current_trans = row[0]
            current_set = []
            current_set.append(row[2])
        else:
            current_set.append(row[2])
    transactions.append([current_trans, current_set])
    return transactions


def compute_support(products, trans, min_occurence):
    # Write your code here to create a list called support.
    # Each element in this will itself be a list like ['cheese', 'egg', 4]
    # Thus, for every unique pair of items such as 'cheese' and 'egg',
    # count the number of transactions that have both of these.
    # If you report for 'cheese' and 'egg', do not also report 'egg' and cheese'
    support = []

    return support

def sort_support(support):
    # We've written this to sort. 
    sorted_support = sorted(support, key=lambda x:x[2], reverse=True)
    return sorted_support