recommendations.py

# David Robertson
# CS1P Lab Exam 2

import string, operator

def get_books(filename='books.txt'):
    books_file = open(filename, 'r')
    books_file_lines = books_file.readlines()
    books_file.close()

    books = map(lambda line: line.strip().split(','), books_file_lines)

    return books

def book_by_index(books, index):
    book = books[index]
    book_string = book[0] + ': ' + book[1]
    return book_string


def get_ratings(filename='ratings.txt'):
    ratings_file = open(filename, 'r')
    ratings_file_lines = ratings_file.readlines()
    ratings_file.close()
    ratings = []
    i = 0
    while i < len(ratings_file_lines):
        name        = ratings_file_lines[i].strip()
        rating_data = map(int, ratings_file_lines[i+1].split()) 
        rating = [name, rating_data]
        ratings.append(rating)
        i += 2
    return ratings

def ratings_by_user(ratings, username):
    filtered = filter(lambda rating: rating[0] == username, ratings)
    if len(filtered) == 0:
        return None
    else:
        return filtered[0][1]

    



def dot_product(a, b):
    return reduce(operator.add, map(operator.mul, a, b))
    
def get_similarities(ratings, user_ratings):
    return map(lambda rating: [rating[0], dot_product(rating[1], user_ratings)], ratings)

def get_recommendations(ratings, user_ratings, targetcount):
    similarities = get_similarities(ratings, user_ratings)
    similarities.sort(key=lambda x: x[1])

    recommendations = []
    seen = []
    
    count = 0
    while count < targetcount and len(similarities) > 0:
        similar_user = similarities.pop()
        similar_user_name    = similar_user[0]
        similar_user_ratings = ratings_by_user(ratings, similar_user_name)
        user_recs = []
        i = 0
        while i < len(similar_user_ratings):
            sur = similar_user_ratings[i]
            ur  = user_ratings[i]
            if ur == 0 and sur > 3 and i not in seen:
                seen.append(i)
                user_recs.append([i, sur])
            i = i + 1
        user_recs.sort(key=lambda x: x[1], reverse=True)
        l = len(user_recs)
        left = targetcount-count
        if l > left:
            user_recs = user_recs[0:left]
            l = left
        if l != 0:            
            user_recs = [similar_user_name, map(lambda x: x[0], user_recs)]
            recommendations.append(user_recs)
            count = count + l
        
    return recommendations


def format_recommendations(recommendations, username, books):
    output = []
    output.append('Book recommendations for ' + username + ':')
    for user_rec in recommendations:
        user, urec_books = user_rec
        output.append('')
        output.append('[Recommended by ' + user + ']')
        for book in urec_books:
            output.append('  * ' + book_by_index(books, book))
    output.append('')
    output_string = string.join(output, sep='\n')
    return output_string

def save_to_file(data, filename='output.txt'):
    fh = open(filename, 'w')
    fh.write(data)
    fh.close()


def get_user(ratings, books):
    username = raw_input('Username: ')
    user_ratings = ratings_by_user(ratings, username)
    if user_ratings is None:
        print 'TODO: input ratings'

    n_recs = raw_input('Number of recommendations: ')
    n_recs = int(n_recs)
    

    return username, user_ratings, n_recs
    
def main():
    books   = get_books()
    ratings = get_ratings()

    username, user_ratings, n_recs = get_user(ratings, books)

    recommendations = get_recommendations(ratings, user_ratings, n_recs)

    output = format_recommendations(recommendations, username, books)
    print output
    save_to_file(output)
    

main()