def recommend_movies(movie_query, k_recommendations):
raw_movies_data = []
with open('movies_recommendation_data.csv', 'r') as md:
# Discard the first line (headings)
next(md)
# Read the data into memory
for line in md.readlines():
data_row = line.strip().split(',')
raw_movies_data.append(data_row)
# Prepare the data for use in the knn algorithm by picking
# the relevant columns and converting the numeric columns
# to numbers since they were read in as strings
movies_recommendation_data = []
for row in raw_movies_data:
data_row = list(map(float, row[2:]))
movies_recommendation_data.append(data_row)
# Use the KNN algorithm to get the 5 movies that are most
# similar to The Post.
recommendation_indices, _ = knn(movies_recommendation_data,
movie_query,
k=k_recommendations,
distance_fn=euclidean_distance,
choice_fn=lambda x: None)
movie_recommendations = []
for _, index in recommendation_indices:
movie_recommendations.append(raw_movies_data[index])
return movie_recommendations
def FindBestwithKNN(restaurant_query, k_recommendations, dataset):
recommendation_indices, _ = knn(dataset,
restaurant_query,
k=k_recommendations,
distance_fn=euclidean_distance,
choice_fn=lambda x: None)
restaurant_recommendations = []
for _, index in recommendation_indices:
restaurant_recommendations.append(dataset[index])
return restaurant_recommendations
def recommend_movies(query,num_of_recommendation):
raw_movies_data = []
with open('movies_recommendation_data.csv','r') as md:
# Discard the first line which is the heading
next(md)
for line in md.readlines():
data_row = line.strip().split(',')
raw_movies_data.append(data_row)
movies_recommmendation_data = []
for row in raw_movies_data:
data_row = list(map(float,row[2:]))
movies_recommmendation_data.append(data_row)
# Use knn to get the k movie recommendation
recommend_indices,_ = knn(movies_recommmendation_data,query,num_of_recommendation,euclidean_distance,lambda x: None)
movie_recommendations = []
for _,index in recommend_indices:
movie_recommendations.append(raw_movies_data[index])
return movie_recommendations
def recommend_movies(movie_query, k_recommendations):
raw_movies_data = []
# open file, read only
with open('movies_recommendation_data.csv', 'r') as md:
next(md) # skip the 'first line'
# read data into memory
for line in md:
data_row = line.strip().split(',')
# debug
print("\n")
print("data_row", data_row)
raw_movies_data.append(data_row)
# debug
print("\n\n")
print("raw_movies_data", raw_movies_data)
# prepare the data, so it can be use by the 'KNN'
# need to convert data in 'each column' into numbers
# because it was 'read in' as a string
movies_recommendation_data = []
for row in raw_movies_data:
# row[2:], get data 'starting' from 'row 2'
# probably don't need ' list() '
data_row = list(map(float, row[2:]))
movies_recommendation_data.append(data_row)
# debug
print("\n\n")
print("movies_recommendation_data", movies_recommendation_data)
print("\n")
print("type movies_recommendation_data", type(movies_recommendation_data) )
# use KNN algorithm to get '5 similar movies' to 'query_movie'
recommendation_indices, jim = knn(
movies_recommendation_data, movie_query, k=k_recommendations,
distance_fn=euclidean_distance, choice_fn=lambda x: None
)
# debug
print("\n\n")
print("recommendation_indices", recommendation_indices)
movie_recommendations = []
for _, index in recommendation_indices:
# debug
print("\n\n")
print("_", _)
print("index", index)
movie_recommendations.append(raw_movies_data[index])
# debug
print("\n\n")
print("movie_recommendations", movie_recommendations)
return movie_recommendations