def represent(id):
data = prefutils.data_from_id(id)
bin_array = prefutils.data_to_bin_array(data, dim)
#print(bin_array)
constraints = []
for i in range(2**dim - 1):
constraints.append([y-x for x,y in zip(bin_array[i],bin_array[i+1])])
#constraints.append(data[i]- data[i+1])
#print(constraints)
b = [-1] * (2**dim -1)
c = [1] * dim
res = linprog(c, constraints, b) #, options={"disp": True})
#print(res)
if res['success']:
return res['x']
else:
return []
# let's count primitive pairs that appear in CSPs
# I am wondering whether using lists of primitive pairs will be more efficient
# that listing the top half of the CSP
# there is some savings, but I think it is more efficient to store a CSP in terms of its
# n parents and the central flip.
dim = 6
id_list = prefutils.get_id_list(dim)
data_list = []
for id in id_list:
data_list.append(prefutils.data_from_id(id))
#print(data_list)
count_list = [0] * 2**dim
for data in data_list:
count = 0
for i in range(len(data)-1):
if data[i] + data[i+1] == data[i] ^ data[i+1]:
#print('\t', data[i], data[i+1])
count = count + 1
#print(id, count)
count_list[count] = count_list[count] + 1
# corresponds to an A (in order) or a B (out of order)
for s in set_list:
x = [0] * dim
x[s[0]] = 1
y = [0] * dim
y[s[1]] = 1
y[s[2]] = 1
pair_dict[s] = (prefutils.bin_array_to_decimal(x),
prefutils.bin_array_to_decimal(y))
type_list = []
for pref in pref_list:
data = prefutils.data_from_id(pref[0])
type = []
for s in set_list:
pair = pair_dict[s]
if (data.index(pair[0]) > data.index(pair[1])):
t = 'B'
else:
t = 'A'
type.append(t)
#print(type)
type_list.append(type)
#cur.execute('SELECT id FROM Pref_Six WHERE name IS NULL LIMIT 5000')
cur.execute('SELECT id FROM Pref_Five')
id_list = []
for (id,) in cur:
id_list.append(id)
count = [0] * dim
middle_dic = {}
for id in id_list:
data = prefutils.data_from_id(id)
#print(data)
middle_num = data[2**(dim-1)]
middle_data = prefutils.decimal_to_bin_array(data[2**(dim-1)], dim)
total = sum(middle_data)
# debug: looking at middle entries with exactly 2 ones.
# where do they appear?
if total == 2:
#print(prefutils.data_to_bin_array(data, dim))
print(middle_data)
if middle_num in middle_dic:
middle_dic[middle_num] = middle_dic[middle_num]+1
else:
middle_dic[middle_num] = 1
def run():
dim = 4
conn = mysql.connector.connect(host='localhost',
database='mysql',
user='******',
password='******')
cur = conn.cursor(buffered=True)
#name_dic = name_dic4a
# populate the name_dic
# name_dic = {}
# cur.execute('SELECT id, name from Pref_Five')
# for (id, name) in cur:
# name_dic[id] = name
keep_running = True
while (keep_running):
# cur.execute('SELECT id FROM Pref_Six WHERE name IS NULL LIMIT 5000')
cur.execute('SELECT id FROM Pref_Five')
id_list = []
name_list = []
for (id, ) in cur:
id_list.append(id)
# keep_running = len(id_list) > 0
keep_running = False
if (keep_running):
print('processed starting with', id_list[0])
for id in id_list:
# for id in [ id_list[0], id_list[1]]:
data = prefutils.data_from_id(id)
# print('++++++++++++ handling id', id)
address = generate_address(data, dim)
print(id, address)
regen_id = regenerate_id(address, dim)
# print(address, id)
# print(data)
#regen_data = [prefutils.bin_array_to_decimal(x) for x in regen_bin_data]
#regen_id = prefutils.halfdata_to_id(regen_data)
num_errors = 0
if not regen_id == id:
print('error id and regen_id dont match', id, regen_id)
num_errors = num_errors + 1
# for a,b in zip(data_bin, regen_bin_data):
# print(a,b)
# print('name array', address)
name = '(' + ','.join(str(x) for x in address) + ')'
name_list.append(name)
# for id, name in zip(id_list, name_list):
# print(id, name)
# print(name_list)
print(len(name_list))
print(len(list(set(name_list))))
print('num errors', num_errors)
# print('adding to DB')
param_list = [(name, id) for name, id in zip(name_list, id_list)]
# cur.executemany('UPDATE Pref_Five SET name=%s WHERE id=%s', param_list)
# cur.executemany('UPDATE Pref_Six SET name=%s WHERE id=%s', param_list)
# conn.commit()
cur.close()