def count_local_triangle(tbl_name, conn):
""" input is a matrix """
tol = 0.1
b = 'b'
lim = 10
cur = conn.cursor()
tn = tbl_name
print "create matrix..."
create_vector_or_matrix(b, conn)
print "Counting dimension..."
calc_dim_query = "select max(row), max(col) from %s" % (tn)
cur.execute(calc_dim_query)
p = cur.fetchone()
dim = max(p) + 1
print "dimension is %s" % dim
appro = min(dim, lim)
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" %
(b, i, 0, 1.0 / float(dim)))
print "init b..."
lanczos(tn, b, dim, appro, conn)
ltriangle = "task7_local_triangle_count"
drop_if_exists(ltriangle, conn)
cur = conn.cursor()
cur.execute("create table %s (nid int, cnt int)" % ltriangle)
print "calculating the local triangle count for each node"
cur.execute(
"insert into %s select U.row, round(sum(power(U.value, 2.0) * power(V.value, 3.0) / 2.0)) from eigenvec U, eigenval V where U.col = V.row and V.row = V.col group by U.row"
% (ltriangle))
def test_uci_gamma(self):
"""http://konect.uni-koblenz.de/networks/ucidata-gama"""
data_file = "data/ucidata-gama.txt"
gfile = "task5_ucigama"
cur = self.conn.cursor()
drop_if_exists(gfile, self.conn)
cur.execute("create table %s(row int, col int, value real DEFAULT 1)" %
(gfile))
cur.execute(
"create table tmppppp(row int, col int, value real DEFAULT 1)")
cur.copy_from(open(data_file),
"tmppppp",
columns=('row', 'col', 'value'),
sep="\t")
cur.execute("insert into %s select row, col from tmppppp" % gfile)
drop_if_exists("tmppppp", self.conn)
cur.execute("update %s set col=col-1, row=row-1" % gfile)
self.conn.commit()
print "Uci gama builded ..."
reverse_matrix(gfile, self.conn)
b = 'b'
create_vector_or_matrix(b, self.conn)
cur.execute("select max(col) from %s" % gfile)
dim = cur.fetchone()[0] + 1
print "Dimension is %s" % dim
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" %
(b, i, 0, 1.0 / float(dim)))
lanczos(gfile, b, dim, 10, self.conn)
drop_if_exists(b, self.conn)
def count_triangle(tbl_name, conn):
""" input is a matrix """
tol = 0.1
b = 'b'
lim = 10
cur = conn.cursor()
tn = tbl_name
print "create matrix..."
create_vector_or_matrix(b, conn)
print "Counting dimension..."
calc_dim_query = "select max(row), max(col) from %s" % (tn)
cur.execute(calc_dim_query)
p = cur.fetchone()
dim = max(p) + 1
print "dimension is %s" % dim
appro = min(dim, lim);
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" % (b, i, 0, 1.0 / float(dim)))
print "init b..."
lanczos(tn, b, dim, appro, conn)
cur.execute("select power(value,3.0) from eigenval where row = col order by value desc limit %s" % appro)
r = cur.fetchall()
print r
s = 0.0001
for i in range(len(r)):
if r[i][0] < 0 or abs(r[i][0]) / s < tol:
break
s += r[i][0]
drop_if_exists(tn, conn)
drop_if_exists(b, conn)
# drop_if_exists("eigenval", conn)
# drop_if_exists("eigenvec", conn)
return s / 6.0
def count_local_triangle(tbl_name, conn):
""" input is a matrix """
tol = 0.1
b = 'b'
lim = 10
cur = conn.cursor()
tn = tbl_name
print "create matrix..."
create_vector_or_matrix(b, conn)
print "Counting dimension..."
calc_dim_query = "select max(row), max(col) from %s" % (tn)
cur.execute(calc_dim_query)
p = cur.fetchone()
dim = max(p) + 1
print "dimension is %s" % dim
appro = min(dim, lim);
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" % (b, i, 0, 1.0 / float(dim)))
print "init b..."
lanczos(tn, b, dim, appro, conn)
ltriangle = "task7_local_triangle_count"
drop_if_exists(ltriangle, conn)
cur = conn.cursor()
cur.execute("create table %s (nid int, cnt int)" % ltriangle)
print "calculating the local triangle count for each node"
cur.execute("insert into %s select U.row, round(sum(power(U.value, 2.0) * power(V.value, 3.0) / 2.0)) from eigenvec U, eigenval V where U.col = V.row and V.row = V.col group by U.row" % (ltriangle))
def test_lanczos(conn):
b = 'b'
v = 'v'
create_vector_or_matrix(v, conn)
clear_table("v", conn)
# random_square_matrix(v, dim, conn);
cur = conn.cursor()
f = open("data/matrix_data.txt")
data = [line.strip().split(" ") for line in f]
for i in range(len(data)):
for j in range(len(data)):
cur.execute("insert into %s values (%s, %s, %s)" % (v, i, j, data[i][j]))
dim = len(data)
create_vector_or_matrix(b, conn)
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" % (b, i, 0, 1.0 / float(dim)))
lanczos(v, b, dim, dim, conn)
def test_lanczos(conn):
b = 'b'
v = 'v'
create_vector_or_matrix(v, conn)
clear_table("v", conn)
# random_square_matrix(v, dim, conn);
cur = conn.cursor()
f = open("data/matrix_data.txt")
data = [line.strip().split(" ") for line in f]
for i in range(len(data)):
for j in range(len(data)):
cur.execute("insert into %s values (%s, %s, %s)" %
(v, i, j, data[i][j]))
dim = len(data)
create_vector_or_matrix(b, conn)
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" %
(b, i, 0, 1.0 / float(dim)))
lanczos(v, b, dim, dim, conn)
def test_ucidata(self):
data_file = "data/ucidata-zachary.txt"
gfile = "task5_ucidata"
cur = self.conn.cursor()
drop_if_exists(gfile, self.conn)
cur.execute("create table %s(row int, col int, value real DEFAULT 1)" % (gfile))
cur.copy_from(open(data_file), gfile, columns=('row', 'col'), sep=" ")
cur.execute("update %s set col=col-1, row=row-1" % gfile)
self.conn.commit()
print "Ucidata builded ..."
reverse_matrix(gfile, self.conn)
b = 'b'
create_vector_or_matrix(b, self.conn)
cur.execute("select max(col) from %s" % gfile)
dim = cur.fetchone()[0] + 1
print "Dimension is %s" % dim
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" % (b, i, 0, 1.0 / float(dim)))
lanczos(gfile, b, dim, 20, self.conn)
drop_if_exists(b, self.conn)
def test_youtube(self):
"""http://snap.stanford.edu/data/com-Youtube.html"""
data_file = "data/com-youtube.ungraph.txt"
gfile = "task5_youtube"
cur = self.conn.cursor()
drop_if_exists(gfile, self.conn)
cur.execute("create table %s(row int, col int, value real DEFAULT 1)" % (gfile))
cur.copy_from(open(data_file), gfile, columns=('row', 'col'), sep="\t")
cur.execute("update %s set col=col-1, row=row-1" % gfile)
self.conn.commit()
print "Youtube builded ..."
reverse_matrix(gfile, self.conn)
b = 'b'
create_vector_or_matrix(b, self.conn)
cur.execute("select max(col) from %s" % gfile)
dim = cur.fetchone()[0] + 1
print "Dimension is %s" % dim
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" % (b, i, 0, 1.0 / float(dim)))
lanczos(gfile, b, dim, 20, self.conn)
drop_if_exists(b, self.conn)
def test_ucidata(self):
data_file = "data/ucidata-zachary.txt"
gfile = "task5_ucidata"
cur = self.conn.cursor()
drop_if_exists(gfile, self.conn)
cur.execute("create table %s(row int, col int, value real DEFAULT 1)" %
(gfile))
cur.copy_from(open(data_file), gfile, columns=('row', 'col'), sep=" ")
cur.execute("update %s set col=col-1, row=row-1" % gfile)
self.conn.commit()
print "Ucidata builded ..."
reverse_matrix(gfile, self.conn)
b = 'b'
create_vector_or_matrix(b, self.conn)
cur.execute("select max(col) from %s" % gfile)
dim = cur.fetchone()[0] + 1
print "Dimension is %s" % dim
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" %
(b, i, 0, 1.0 / float(dim)))
lanczos(gfile, b, dim, 20, self.conn)
drop_if_exists(b, self.conn)
def test_youtube(self):
"""http://snap.stanford.edu/data/com-Youtube.html"""
data_file = "data/com-youtube.ungraph.txt"
gfile = "task5_youtube"
cur = self.conn.cursor()
drop_if_exists(gfile, self.conn)
cur.execute("create table %s(row int, col int, value real DEFAULT 1)" %
(gfile))
cur.copy_from(open(data_file), gfile, columns=('row', 'col'), sep="\t")
cur.execute("update %s set col=col-1, row=row-1" % gfile)
self.conn.commit()
print "Youtube builded ..."
reverse_matrix(gfile, self.conn)
b = 'b'
create_vector_or_matrix(b, self.conn)
cur.execute("select max(col) from %s" % gfile)
dim = cur.fetchone()[0] + 1
print "Dimension is %s" % dim
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" %
(b, i, 0, 1.0 / float(dim)))
lanczos(gfile, b, dim, 20, self.conn)
drop_if_exists(b, self.conn)
def test_uci_gamma(self):
"""http://konect.uni-koblenz.de/networks/ucidata-gama"""
data_file = "data/ucidata-gama.txt"
gfile = "task5_ucigama"
cur = self.conn.cursor()
drop_if_exists(gfile, self.conn)
cur.execute("create table %s(row int, col int, value real DEFAULT 1)" % (gfile))
cur.execute("create table tmppppp(row int, col int, value real DEFAULT 1)")
cur.copy_from(open(data_file), "tmppppp", columns=('row', 'col', 'value'), sep="\t")
cur.execute("insert into %s select row, col from tmppppp" % gfile)
drop_if_exists("tmppppp", self.conn)
cur.execute("update %s set col=col-1, row=row-1" % gfile)
self.conn.commit()
print "Uci gama builded ..."
reverse_matrix(gfile, self.conn)
b = 'b'
create_vector_or_matrix(b, self.conn)
cur.execute("select max(col) from %s" % gfile)
dim = cur.fetchone()[0] + 1
print "Dimension is %s" % dim
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" % (b, i, 0, 1.0 / float(dim)))
lanczos(gfile, b, dim, 10, self.conn)
drop_if_exists(b, self.conn)
def count_triangle(tbl_name, conn):
""" input is a matrix """
tol = 0.1
b = 'b'
lim = 10
cur = conn.cursor()
tn = tbl_name
print "create matrix..."
create_vector_or_matrix(b, conn)
print "Counting dimension..."
calc_dim_query = "select max(row), max(col) from %s" % (tn)
cur.execute(calc_dim_query)
p = cur.fetchone()
dim = max(p) + 1
print "dimension is %s" % dim
appro = min(dim, lim)
for i in range(dim):
cur.execute("insert into %s values (%s, %s, %s)" %
(b, i, 0, 1.0 / float(dim)))
print "init b..."
lanczos(tn, b, dim, appro, conn)
cur.execute(
"select power(value,3.0) from eigenval where row = col order by value desc limit %s"
% appro)
r = cur.fetchall()
print r
s = 0.0001
for i in range(len(r)):
if r[i][0] < 0 or abs(r[i][0]) / s < tol:
break
s += r[i][0]
drop_if_exists(tn, conn)
drop_if_exists(b, conn)
# drop_if_exists("eigenval", conn)
# drop_if_exists("eigenvec", conn)
return s / 6.0