def outputSparseMatrixStats(spars_matrix, dict_item_id_to_users):
"""
Analyzing for Matrix
:param spars_matrix:
:param dict_item_id_to_users:
"""
ecc_item_sim_nonzero = spars_matrix.nonzero()
dict_names=create_dict_names()
dict_item_ecc=create_dict_ecc()
i_item_index = ecc_item_sim_nonzero[0]
j_item_index = ecc_item_sim_nonzero[1]
results_to_print=list()
for index in range(len(i_item_index)):
results_to_print.append([i_item_index[index], j_item_index[index], spars_matrix[i_item_index[index], j_item_index[index]]])
results_to_print=sorted(results_to_print, key=itemgetter(2))
for item_pair in results_to_print[-100:]:
print(50*"#")
item_names_print([item_pair[0],item_pair[1]],dict_names,dict_item_ecc)
print("THEIR VALUE IS:",item_pair[2])
a=set(dict_item_id_to_users[int(item_pair[0])])
b=set(dict_item_id_to_users[int(item_pair[1])])
intersection= a&b
print("Intersection Length:",len(intersection))
print(50*"#")
def print_maxes(mat):
"""
prints variouse informations about the martrix
:param mat: matrix to print
"""
u_to_likes = load_or_create("/Matrix/UserIdToLikes.matrix",
create_matrix_user_likes)
dict_names = load_or_create('/DICT/MovieIdToName.dict', create_dict_names)
dict_ecc = load_or_create('/DICT/MovieIdToItemEccentricity.dict',
create_dict_ecc)
user_to_ecc = load_or_create('/DICT/UserIdToUserEccentricity.dict',
createDictUserIdToUserEccentricity)
dict_userid_to_moviesliked = load_or_create(
'/DICT/UserIdToLikedMovies.dict', create_dict_user_id_to_liked_items)
dict_userid_to_recommends = dict()
print("STARTING ECC CALC")
recommends = []
for i in range(int(mat.shape[0] * 0.5)):
row = mat.getrow(i)
if len(row.nonzero()[0]) != 0:
# print(u_to_likes.getrow(i).nonzero()[1])
if len(u_to_likes.getrow(i).nonzero()[1]) < 10 and user_to_ecc[
i + 1] > 0:
# print("Amount of recommends:",len(row.nonzero()[0]))
row = row.toarray()[0].tolist()
max_val = max(val for val in row
if str(row.index(val) +
1) not in dict_userid_to_moviesliked[i +
1])
print(
'SUM is:',
sum(val for val in row
if str(row.index(val) +
1) not in dict_userid_to_moviesliked[i + 1]))
print('SUM with all is:', sum(val for val in row))
index_max = row.index(max_val) + 1
recommends.append([
max_val,
row.index(max_val) + 1, i + 1,
[i + 1 for i in u_to_likes.getrow(i).nonzero()[1]],
[row.index(max_val) + 1], user_to_ecc[i + 1]
])
recommends = sorted(recommends, key=itemgetter(0))
for i in recommends[-100:]:
print("MAX id:", i[1])
print("MAX val:", i[0])
print("Users ECC:", i[5])
print("for user:"******"MOVIES HE ALREADY LIKED", 50 * "=")
item_names_print(i[3], dict_names, dict_ecc)
print("Movie Well recommend:" + 50 * '*')
item_names_print(i[4], dict_names, dict_ecc)
print(50 * "#")
def print_max_similar():
dict_names = load_or_create('/DICT/MovieIdToName.dict', create_dict_names)
dict_ecc = load_or_create('/DICT/MovieIdToItemEccentricity.dict',
create_dict_ecc)
mat = load_or_create("/Matrix/ItemSimilarityEccentricity.matrix",
create_matrix_item_similarity)
maxes = []
for i in range(int(mat.shape[1] * 0.1)):
row = mat.getrow(i)
if len(row) > 0:
max_curr = max(i for i in row)
maxes.append([i + 1, row.index(max_curr) + 1, max_curr])
maxes = sorted(maxes, key=itemgetter(2))
for i, j, val in maxes:
print("Val is:", val)
item_names_print([i, j], dict_names, dict_ecc)
def print_maxes():
u_to_likes = load_or_create("/Matrix/UserIdToLikes.matrix",
create_matrix_user_likes)
dict_names = load_or_create('/DICT/MovieIdToName.dict', create_dict_names)
dict_ecc = load_or_create('/DICT/MovieIdToItemEccentricity.dict',
create_dict_ecc)
mat = get_recommendation_matrix()
user_to_ecc = load_or_create('/DICT/UserIdToUserEccentricity.dict',
createDictUserIdToUserEccentricity)
dict_userid_to_recommends = dict()
recommends = []
for i in range(int(mat.shape[0] * 0.5)):
row = mat.getrow(i)
if len(row.nonzero()[0]) != 0:
# print(u_to_likes.getrow(i).nonzero()[1])
if len(u_to_likes.getrow(i).nonzero()[1]) < 10 and user_to_ecc[
i + 1] > 0:
# print("Amount of recommends:",len(row.nonzero()[0]))
row = row.toarray()[0].tolist()
max_val = max(val for val in row)
print('SUM with all is:', sum(val for val in row))
recommends.append([
max_val,
row.index(max_val) + 1, i + 1,
[i + 1 for i in u_to_likes.getrow(i).nonzero()[1]],
[row.index(max_val) + 1], user_to_ecc[i + 1]
])
recommends = sorted(recommends, key=itemgetter(0))
for i in recommends[-100:]:
print("MAX id:", i[1])
print("MAX val:", i[0])
print("Users ECC:", i[5])
print("for user:"******"MOVIES HE ALREADY LIKED", 50 * "=")
item_names_print(i[3], dict_names, dict_ecc)
print("Movie Well recommend:" + 50 * '*')
item_names_print(i[4], dict_names, dict_ecc)
print(50 * "#")