def int_to_Hilbert_test(n, nD, doPrint=True):
tups_per_line = 2**(int(log(80 / tuple_print_width(nD), 2)))
linear_size = int((n + .5)**(1. / nD))
isCube = linear_size**nD == n
if not isCube:
print "(This is not a complete cube:)"
visited = {}
for i in range(n):
pt = tuple(int_to_Hilbert(i, nD))
if doPrint: print pt,
visited[pt] = True # Add pt to "visited" dictionary.
j = Hilbert_to_int(pt)
if j != i:
print "BZZT, decodes to", j,
if i > 0: # Check that we have stepped one unit, along one dim:
nDiffs = sum([pt[k] != prev_pt[k] for k in range(nD)])
if nDiffs != 1:
print "BZZT, different along", nDiffs, "dimensions",
maxDiff = max([abs(pt[k] - prev_pt[k]) for k in range(nD)])
if maxDiff != 1:
print "BZZT, max difference is", maxDiff,
prev_pt = pt
if (i + 1) % tups_per_line == 0:
if doPrint: print
if isCube:
check_visited(visited, [], linear_size, nD)
def int_to_Hilbert_test( n, nD, doPrint=True ):
tups_per_line = 2**( int( log( 80 / tuple_print_width( nD ), 2 ) ) )
linear_size = int( ( n + .5 ) ** (1./nD) )
isCube = linear_size ** nD == n
if not isCube:
print "(This is not a complete cube:)"
visited = {}
for i in range( n ):
pt = tuple( int_to_Hilbert( i, nD ) )
if doPrint: print pt,
visited[ pt ] = True # Add pt to "visited" dictionary.
j = Hilbert_to_int( pt )
if j != i:
print "BZZT, decodes to", j,
if i > 0: # Check that we have stepped one unit, along one dim:
nDiffs = sum( [ pt[k] != prev_pt[k] for k in range( nD ) ] )
if nDiffs != 1:
print "BZZT, different along", nDiffs, "dimensions",
maxDiff = max( [ abs(pt[k] - prev_pt[k]) for k in range( nD ) ] )
if maxDiff != 1:
print "BZZT, max difference is", maxDiff,
prev_pt = pt
if ( i + 1 ) % tups_per_line == 0:
if doPrint: print
if isCube:
check_visited( visited, [], linear_size, nD )
def get(self, h):
h = long(h)
woeid, ts = hilbert.int_to_Hilbert(h, 2)
self.api_ok(self.generate_response(woeid, ts, h))
return
def get(self, h):
h = long(h)
addr_int, ts = hilbert.int_to_Hilbert(h, 2)
addr = "%d.%d.%d.%d" % ((addr_int >> 24)&0xFF, (addr_int >> 16)&0xFF, (addr_int >> 8)&0xFF, addr_int&0xFF)
self.api_ok(self.generate_response(addr, ts, h))
return
def get(self, h):
h = long(h)
x, y, z = hilbert.int_to_Hilbert(h, 3)
x = x / self.factor
y = y / self.factor
x -= 180.
y -= 90.
self.api_ok(self.generate_response(x, y, z, h))
return
# If needed...
def generateface2picsmapping(minimum_faces_per_person=1):
lfw_people = fetch_lfw_people(min_faces_per_person=minimum_faces_per_person, resize=0.4)
n_samples, h, w = lfw_people.images.shape
X, y, target_names = lfw_people.data, lfw_people.target, lfw_people.target_names
n_examples, n_features = X.shape
face2pics = []
print(max(y))
for i in range((max(y)+1)):
face2pics.append([target_names[i],[] ])
for i in range(len(y)):
face2pics[y[i]][1].append(i)
return face2pics
if __name__ == "__main__":
(X, X_nmf, y, nmf_components, min_faces_per_person, face2pics) = pickle.load(open( "nmfdata.p", "rb" ) )
n_examples, n_features = X_nmf.shape
print(X_nmf[0])
i = Hilbert_to_int( float2IntVector(X_nmf[0] ))
print(i)
v2 = int2FloatVector(int_to_Hilbert( i, nD=n_features ))
print(v2)
i2 = Hilbert_to_int( float2IntVector(v2 ))
print(i2)
print('loss')
print(i2-i)
# child_start_end_tests()
int_to_Hilbert_test( 8, 1 )
print
int_to_Hilbert_test( 64, 2 )
print
int_to_Hilbert_test( 64, 3 )
print
int_to_Hilbert_test( 4096, 3, doPrint=False ) # nChunks > nD
print
int_to_Hilbert_test( 4096, 4, doPrint=False )
print
k = 10**12 # definitely a long
for nD in [ 2, 3, 4, 5 ]:
# Decode point ( k, 0, 0... ) to an int...how many digits?:
pt = [ 0 ] * nD
pt[0] = k
x = Hilbert_to_int( pt )
print x, log( x, 10 )
# Double-check:
pt2 = list( int_to_Hilbert( x, nD ) )
if pt2 != pt:
print "BZZT! Encodes to ", pt2
# Encode k to an nD point:
pt = int_to_Hilbert( k, nD )
print pt
# Double-check:
y = Hilbert_to_int( pt )
if y != k:
print "BZZT! Decodes to", y
print int_to_Hilbert_test(64, 3)
print "int_to_Hilbert_test( 81, 4)"
print int_to_Hilbert_test(81, 4)
print "weird"
print int_to_Hilbert_test(256, 4)
print "int_to_Hilbert_test( 4096, 3, doPrint=False ) # nChunks > nD"
print int_to_Hilbert_test(4096, 3, doPrint=False) # nChunks > nD
print "int_to_Hilbert_test( 4096, 4, doPrint=False )"
print int_to_Hilbert_test(4096, 4, doPrint=False)
k = 10**12 + 0.2 # definitely a long
for nD in [2, 3, 4, 5, 20, 40, 2000]:
# Decode point ( k, 0, 0... ) to an int...how many digits?:
pt = [0] * nD
pt[0] = k
x = Hilbert_to_int(pt)
print x, log(x, 10)
# Double-check:
pt2 = list(int_to_Hilbert(x, nD))
if pt2 != pt:
print "BZZT! Encodes to ", pt2
# Encode k to an nD point:
pt = int_to_Hilbert(k, nD)
print pt
# Double-check:
y = Hilbert_to_int(pt)
if y != k:
print "BZZT! Decodes to", y
# If needed...
def generateface2picsmapping(minimum_faces_per_person=1):
lfw_people = fetch_lfw_people(
min_faces_per_person=minimum_faces_per_person, resize=0.4)
n_samples, h, w = lfw_people.images.shape
X, y, target_names = lfw_people.data, lfw_people.target, lfw_people.target_names
n_examples, n_features = X.shape
face2pics = []
print(max(y))
for i in range((max(y) + 1)):
face2pics.append([target_names[i], []])
for i in range(len(y)):
face2pics[y[i]][1].append(i)
return face2pics
if __name__ == "__main__":
(X, X_nmf, y, nmf_components, min_faces_per_person,
face2pics) = pickle.load(open("nmfdata.p", "rb"))
n_examples, n_features = X_nmf.shape
print(X_nmf[0])
i = Hilbert_to_int(float2IntVector(X_nmf[0]))
print(i)
v2 = int2FloatVector(int_to_Hilbert(i, nD=n_features))
print(v2)
i2 = Hilbert_to_int(float2IntVector(v2))
print(i2)
print('loss')
print(i2 - i)
def mk_int(i):
(x, y, z)= hilbert.int_to_Hilbert(i, 3)
rsp = flask.jsonify(x=x, y=y, z=z, i=i)
return rsp