def makeA(u):
ulistofalldensities = {}
i = 0
usz = set() #set of all equivalence class sizes that we have seen
for fileitem in l: #each file is a dif density
d = zkl.load(fileitem)
c = get_counts(d, u)
ulistofalldensities[densities[i]] = c
i = i + 1
for v in c:
usz.add(v)
print usz
#print ulistofalldensities
#print usz
for density in ulistofalldensities:
for size in usz:
if not size in ulistofalldensities[density]:
ulistofalldensities[density][size] = 0
A = []
for density in densities:
A.append([
ulistofalldensities[density][x]
for x in np.sort(ulistofalldensities[density].keys())
])
return A
def makeA(U):
ulistofallrates = {}
i = 0
usz = set() #set of all rates that we have seen
for fileitem in l: #each file is a dif density
d = zkl.load(fileitem)
c = get_counts(d,U)
ulistofallrates[densities[i]] = c
i = i + 1
for v in c:
usz.add(v)
print usz
for density in ulistofallrates:
for rate in usz:
if not rate in ulistofallrates[density]:
ulistofallrates[density][rate] = 0
A = []
for density in densities:
A.append([ulistofallrates[density][x] for x in np.sort(ulistofallrates[density].keys())])
return A
def makeA(U):
ulistofallrates = {}
i = 0
usz = set() #set of all rates that we have seen
for fileitem in l: #each file is a dif density
d = zkl.load(fileitem)
c = get_counts(d, U)
ulistofallrates[densities[i]] = c
i = i + 1
for v in c:
usz.add(v)
print usz
for density in ulistofallrates:
for rate in usz:
if not rate in ulistofallrates[density]:
ulistofallrates[density][rate] = 0
A = []
for density in densities:
A.append([
ulistofallrates[density][x]
for x in np.sort(ulistofallrates[density].keys())
])
return A
def get_subplot_x_y(density_of_gt_file,undersampling):
d = zkl.load(density_of_gt_file)
x = []
y = []
for i in range(0,100):
g2 = bfutils.undersample(d[i]['gt'],undersampling) #this is H
x.append(traversal.density(g2)) #add the density of H
y.append(d[i]['solutions'][undersampling]['ms']) #add the time
return x,y
def return_only_unique_rg(name):
rg = zkl.load(name)
for item1 in rg:
#delete all isomorphisms from rg
all_isos = find_all_isomorphisms(item1)
for item2 in all_isos:
if item2 in rg and item1 != item2:
rg.remove(item2)
return rg
def get_subplot_x_y(density_of_gt_file, undersampling):
d = zkl.load(density_of_gt_file)
x = []
y = []
for i in range(0, 100):
g2 = bfutils.undersample(d[i]['gt'], undersampling) #this is H
x.append(traversal.density(g2)) #add the density of H
y.append(d[i]['solutions'][undersampling]['ms']) #add the time
return x, y
def make_unique_urg_and_pics(num_nodes):
uurgs = determine_unique_unreachable_graphs('rgs_%s.zkl' % num_nodes,num_nodes)
zkl.save(uurgs,'unique_urgs_%s.zkl' % num_nodes)
print "Finished importing. Now drawing:"
d = zkl.load('unique_urgs_%s.zkl' % num_nodes)
i = 1
for item in d:
draw_codomain_graph(item,'pic_unique_urgs_%s_graph_%d.png' % (num_nodes,i))
print '\n'
print item
i = i + 1
def fastest_g(L):
x = []
y = []
for l in L:
d = zkl.load(l)
for i in range(0,100):
gs = bfutils.call_undersamples(d[i]['gt']) #this helps us determine how far u will go
for u in range(1,len(d[i]['solutions'])+1):
g2 = bfutils.undersample(d[i]['gt'],u) #this is H
x.append(traversal.density(g2)) #add the density of H
y.append(d[i]['solutions'][u]['ms']) #add the time
return x,map(lambda x: x/1000./60., y)
def fastest_g(L):
x = []
y = []
for l in L:
d = zkl.load(l)
for i in range(0, 100):
gs = bfutils.call_undersamples(
d[i]['gt']) #this helps us determine how far u will go
for u in range(1, len(d[i]['solutions']) + 1):
g2 = bfutils.undersample(d[i]['gt'], u) #this is H
x.append(traversal.density(g2)) #add the density of H
y.append(d[i]['solutions'][u]['ms']) #add the time
return x, map(lambda x: x / 1000. / 60., y)
def gen_x_y(L):
x = []
y = []
for l in L:
d = zkl.load(l)
for i in range(0,100):
gs = bfutils.call_undersamples(d[i]['gt']) #this helps us determine how far u will go
for u in range(1,len(d[i]['solutions'])+1):
#for u in range(1,min([len(gs),4])):
g2 = bfutils.undersample(d[i]['gt'],u) #this is H
x.append(traversal.density(g2)) #add the density of H
y.append(d[i]['solutions'][u]['ms']) #add the time
return x,y
def gen_x_y(L):
x = []
y = []
for l in L:
d = zkl.load(l)
for i in range(0, 100):
gs = bfutils.call_undersamples(
d[i]['gt']) #this helps us determine how far u will go
for u in range(1, len(d[i]['solutions']) + 1):
#for u in range(1,min([len(gs),4])):
g2 = bfutils.undersample(d[i]['gt'], u) #this is H
x.append(traversal.density(g2)) #add the density of H
y.append(d[i]['solutions'][u]['ms']) #add the time
return x, y
def determine_unique_unreachable_graphs(name,n): #determine codomain codomain = determine_codomain(n) #SLOW #load up zkl file of all reachable graphs rgs = zkl.load(name) #for every graph in reachable graphs, create all possible forms and remove from codomain for rg in rgs: allforms_graph = find_all_forms_codomain_graph(rg) for item in allforms_graph: while item in codomain: codomain.remove(item) #remove isomorphisms from codomain to get unique codomain for graph1 in codomain: all_isos = find_all_isomorphisms(graph1) for graph2 in all_isos: if graph2 in codomain and graph1 != graph2: codomain.remove(graph2) return codomain
import graphkit as gk
import bfutils as bfu
import numpy as np
import pandas as pd
from matplotlib import pyplot as plt
#
uplimit = 0.25
#
#d05 = zkl.load('leibnitz_nodes_5_samples_2000_noise_0.1_OCE_b_svar_beta_rasl_more.zkl')
#d05 = zkl.load('oranos_nodes_5_samples_2000_noise_0.1_OCE_b_svar_beta_rasl_more.zkl')
#d05 = zkl.load('leibnitz_nodes_6_samples_2000_noise_0.1_OCE_b_svar_beta_rasl.zkl')
d05 = zkl.load(
'leibnitz_nodes_6_samples_2000_noise_0.1_OCE_b_svar_beta_rasl.zkl')
def estOE(d):
gt = d['gt']['graph']
gt = bfu.undersample(gt, 1)
e = gk.OCE(d['estimate'], gt)
N = np.double(len(gk.edgelist(gt))) +\
np.double(len(gk.bedgelist(gt)))
return (e['directed'][0] + e['bidirected'][0]) / N
def estCOE(d):
gt = d['gt']['graph']
gt = bfu.undersample(gt, 1)
e = gk.OCE(d['estimate'], gt)
import StringIO
import scipy
import sys
import os
import igraph
import numpy as np
# local packages
import zickle
import ecj
from ecj import undersample
import testgraphs
reload(testgraphs)
from testgraphs import *
colors = zickle.load('data/colors.zkl') # load the colors
def graph2dict(g):
D = {}
for v in range(0,len(g.vs)):
D[g.vs[v]["label"]] = {}
for u in g.neighbors(v,mode="OUT"):
D[g.vs[v]["label"]][g.vs[u]["label"]] = set([(0,1)])
return D
def paintSCC(g, cm):
nameidx = {}
D = graph2dict(g)
scc = ecj.scc(D)
for i in range(0,len(scc)):
for v in scc[i]:
import gmpy as gmp
import gmpy as gmp
from scipy.misc import comb
import zickle as zkl
import simpleloops as sls
import math
import load_loops
from matplotlib.cbook import flatten
from progressbar import ProgressBar, Percentage, \
Bar, RotatingMarker, ETA, FileTransferSpeed
TOOLSPATH='./tools/'
sys.path.append(os.path.expanduser(TOOLSPATH))
circp = zkl.load('circular_p.zkl')
alloops = load_loops.alloops
import pprint
import bfutils as bfu
import traversal as trv
import graphkit as gk
import comparison as cmp
import simpleloops as sl
def memo(func):
cache = {} # Stored subproblem solutions
@wraps(func) # Make wrap look like func
def wrap(*args): # The memoized wrapper
s = trv.gsig(args[0]) # Signature: just the g
#s = tool.signature(args[0],args[2])# Signature: g and edges
def gettimes(d):
t = [x['ms'] for x in d]
time = map(lambda x: x / 1000. / 60., t)
return time
l = [(0.15, 'leibnitz_nodes_15_density_0.1_newp_.zkl'),
(0.20, 'leibnitz_nodes_20_density_0.1_newp_.zkl'),
(0.25, 'leibnitz_nodes_25_density_0.1_newp_.zkl'),
(0.30, 'leibnitz_nodes_30_density_0.1_newp_.zkl'),
(0.35, 'leibnitz_nodes_35_density_0.1_newp_.zkl')]
fig = pl.figure(figsize=[10, 3])
#Read in data & create total column
d = zkl.load("hooke_nodes_6_g32g1_.zkl") #hooke_nodes_35_newp_.zkl")
densities = [.15, .20, .25, .30, .35]
d = {}
for fname in l:
d[fname[0]] = zkl.load(fname[1])
def get_counts(d):
eqc = [len(x['eq']) for x in d]
keys = np.sort(np.unique(eqc))
c = {}
for k in keys:
c[k] = len(np.where(eqc == k)[0])
return c
import functools
import zickle as zkl
import time, socket
import scipy
import zickle as zkl
import traversal, bfutils, graphkit, unknownrate
import bfutils as bfu
import matplotlib.pyplot as plt
NODES = 6
DENSITY = 0.2
UMAX = 6
REPEATS = 100
d = zkl.load('leibnitz_nodes_6_density_0.2_ra_.zkl')
x = []
y = []
for i in range(0, REPEATS):
gs = bfutils.call_undersamples(
d[i]['gt']) #this helps us determine how far u will go
for u in range(1, min([len(gs), UMAX])):
g2 = bfutils.undersample(d[i]['gt'], u) #this is H
x.append(traversal.density(g2)) #add the density of H
y.append(d[i]['solutions'][u]['ms']) #add the time
print len(x)
print len(y)
fig = plt.figure()
import bfutils as bfu
import numpy as np
import pandas as pd
from matplotlib import pyplot as plt
#
uplimit = 0.25
#
#d05 = zkl.load('leibnitz_nodes_5_samples_2000_noise_0.1_OCE_b_svar_beta_rasl_more.zkl')
#d05 = zkl.load('oranos_nodes_5_samples_2000_noise_0.1_OCE_b_svar_beta_rasl_more.zkl')
#d05 = zkl.load('leibnitz_nodes_6_samples_2000_noise_0.1_OCE_b_svar_beta_rasl.zkl')
d05 = zkl.load('leibnitz_nodes_6_samples_2000_noise_0.1_OCE_b_svar_beta_rasl.zkl')
def estOE(d):
gt= d['gt']['graph']
gt=bfu.undersample(gt,1)
e = gk.OCE(d['estimate'],gt)
N = np.double(len(gk.edgelist(gt))) +\
np.double(len(gk.bedgelist(gt)))
return (e['directed'][0]+e['bidirected'][0])/N
def estCOE(d):
gt= d['gt']['graph']
gt=bfu.undersample(gt,1)
e = gk.OCE(d['estimate'],gt)
n = len(gt)
N = np.double(n**2+(n-1)**2/2.0\
t = [x['ms'] for x in d]
time = map(lambda x: x / 1000. / 60., t)
return time
l = [
'leibnitz_nodes_15_density_0.1_newp_.zkl',
'leibnitz_nodes_20_density_0.1_newp_.zkl',
'leibnitz_nodes_25_density_0.1_newp_.zkl',
'leibnitz_nodes_30_density_0.1_newp_.zkl',
'leibnitz_nodes_35_density_0.1_newp_.zkl'
]
alltimes_new = []
for fname in l:
d = zkl.load(fname)
alltimes_new.append(gettimes(d))
shift = 0.15
wds = 0.3
fliersz = 2
lwd = 1
plt.figure(figsize=[10, 2])
g = sb.boxplot(alltimes_new,
names=map(lambda x: str(int(x * 100)) + "", densities),
widths=wds,
color="Reds",
fliersize=fliersz,
linewidth=lwd,
densities = [.15, 0.2, 0.25, 0.3, 0.35]
def gettimes(d):
t = [x['ms'] for x in d]
time = map(lambda x: x/1000./60., t)
return time
l = ['leibnitz_nodes_15_density_0.1_newp_.zkl',
'leibnitz_nodes_20_density_0.1_newp_.zkl',
'leibnitz_nodes_25_density_0.1_newp_.zkl',
'leibnitz_nodes_30_density_0.1_newp_.zkl',
'leibnitz_nodes_35_density_0.1_newp_.zkl']
alltimes_new = []
for fname in l:
d = zkl.load(fname)
alltimes_new.append(gettimes(d))
shift = 0.15
wds = 0.3
fliersz = 2
lwd = 1
plt.figure(figsize=[10,2])
g = sb.boxplot(alltimes_new,names=map(lambda x: str(int(x*100))+"",
densities),
widths=wds, color="Reds",fliersize=fliersz,
linewidth=lwd,
**{'positions':np.arange(len(densities))+shift,
import zickle as zkl
alloops = zkl.load('allloops.zkl')
import os
import zickle as zkl
DIR_NAME = os.path.dirname(__file__)
ABS_PATH = os.path.abspath(os.path.join(DIR_NAME))
alloops = zkl.load('{}/../data/allloops.zkl'.format(ABS_PATH))
circp = zkl.load('{}/../data/circular_p.zkl'.format(ABS_PATH))
import scipy
import sys
import os
import igraph
import numpy as np
# local packages
import zickle
import ecj
from ecj import undersample
import testgraphs
reload(testgraphs)
from testgraphs import *
colors = zickle.load('data/colors.zkl') # load the colors
def graph2dict(g):
D = {}
for v in range(0, len(g.vs)):
D[g.vs[v]["label"]] = {}
for u in g.neighbors(v, mode="OUT"):
D[g.vs[v]["label"]][g.vs[u]["label"]] = set([(0, 1)])
return D
def paintSCC(g, cm):
nameidx = {}
D = graph2dict(g)
scc = ecj.scc(D)
else:
eqc.append(len(d[i]['solutions'][tempu]['eq']))
keys = np.sort(np.unique(eqc))
c = {}
for k in keys:
c[k] = len(np.where(
eqc == k)[0]) #key is equiv class size and value is frequency
print "u:", U + 1
print "c:", c
print "\n"
return c
#dc is a dictionary where key = u being considered and value = dictionary of equivalence class sizes
#usz is a set of all equivalence class sizes(not frequencies)
d = zkl.load(l)
usz = set()
dc = {}
for u in range(UMAX):
dc[u] = get_counts(d, u + 1)
for v in dc[u]:
usz.add(v)
print usz
#adding all seen equivalence sizes for every density
for u in range(UMAX):
for c in usz:
if not c in dc[u]:
dc[u][c] = 0
#A is all eqc frequencies for all density from bottom to top
def listplot(fname, mname='JJ', stl='', width=5):
l = zkl.load(fname)
l = l[:17*10]
y = min(width,len(l))
x = np.int(np.ceil(len(l)/float(y)))
d2l.matrix_list(l,y,x,R=2, w_gap=1, h_gap=2, mname=mname, stl=stl)
def gettimes(d):
t = [x['ms'] for x in d]
time = map(lambda x: x/1000./60., t)
return time
l = [(0.15, 'leibnitz_nodes_15_density_0.1_newp_.zkl'),
(0.20, 'leibnitz_nodes_20_density_0.1_newp_.zkl'),
(0.25, 'leibnitz_nodes_25_density_0.1_newp_.zkl'),
(0.30, 'leibnitz_nodes_30_density_0.1_newp_.zkl'),
(0.35, 'leibnitz_nodes_35_density_0.1_newp_.zkl')]
fig = pl.figure(figsize=[10,3])
#Read in data & create total column
d = zkl.load("hooke_nodes_6_g32g1_.zkl")#hooke_nodes_35_newp_.zkl")
densities = [.15, .20 ,.25, .30, .35]
d = {}
for fname in l:
d[fname[0]] = zkl.load(fname[1])
def get_counts(d):
eqc = [len(x['eq']) for x in d]
keys = np.sort(np.unique(eqc))
c = {}
for k in keys:
c[k] = len(np.where(eqc == k)[0])
return c
# unique size
usz = set()
if d[i]['solutions'][U]['eq'] == set([-1]): #supercliques
t.append(d[i]['solutions'][U-1]['ms'])
else:
t.append(d[i]['solutions'][U]['ms'])
else: #len of d[i]['solutions') < U
if d[i]['solutions'][len(d[i]['solutions'])]['eq'] == set([-1]): #supercliques
t.append(d[i]['solutions'][len(d[i]['solutions'])-1]['ms'])
else:
t.append(d[i]['solutions'][len(d[i]['solutions'])]['ms'])
#print len(t)
time = map(lambda x: x/1000./60., t)
return time
listofalltimes = []
for fileitem in l:
d = zkl.load(fileitem)
alltimes = []
for u in range(UMAX):
partial_times = gettimes(d,u+1)
alltimes.append(partial_times)
listofalltimes.append(alltimes)
#REFERENCE
#listofalltimes[0] = times for density 25
#listofalltimes[1] = times for density 30
#listofalltimes[2] = times for density 35
#listofalltimes[0][0] = times for density 25 with u = 2
from matplotlib import pyplot as plt
import matplotlib as mpl
import seaborn as sns
import numpy as np
import sys
sys.path.append('/na/home/splis/soft/src/dev/craft/gunfolds/tools/')
sys.path.append('/na/homes/splis/soft/src/dev/tools/stackedBarGraph/')
import zickle as zkl
from stackedBarGraph import StackedBarGrapher
SBG = StackedBarGrapher()
fig = pl.figure(figsize=[10,1.3])
#Read in data & create total column
d = zkl.load("hooke_nodes_6_g32g1_.zkl")#hooke_nodes_35_newp_.zkl")
densities = np.sort(d.keys())
def get_counts(d):
eqc = [len(x['eq']) for x in d]
keys = np.sort(np.unique(eqc))
c = {}
for k in keys:
c[k] = len(np.where(eqc == k)[0])
return c
# unique size
usz = set()
dc = {}
for u in densities:
dc[u] = get_counts(d[u])
#eigenvalues
#pivotal nodes (If X is a pivotal node for Y and Z in a graph in the eqc, does
#this hold for all other graphs in the same eqc?)
#centrality
#groupoids
#todo:
#create ALL H's to make eqcs (so far only 3 nodes...or random H's)
Hs = zkl.load("H_3.zkl")
eqcs = zkl.load("eqcs_for_H_3.zkl")
nonempty_eqcs = []
for eqc in eqcs:
if eqc != set([]):
nonempty_eqcs.append(eqc)
print nonempty_eqcs
# for i in range(len(Hs)):
# H = Hs[i]
# eqc = eqcs[i]
# if eqc != set([-1]):
# print "H: ",determine_all_betweeness_centrality(H)
# print eqc
# for graphstr in eqc:
# graph = comparison.num2CG(graphstr,len(H))
def loadgraphs(fname):
g = zkl.load(fname)
return g
def loadgraphs(fname):
g = zkl.load(fname)
return g
else:
t.append(d[i]['solutions'][U]['ms'])
else: #len of d[i]['solutions') < U
if d[i]['solutions'][len(d[i]['solutions'])]['eq'] == set(
[-1]): #supercliques
t.append(d[i]['solutions'][len(d[i]['solutions']) - 1]['ms'])
else:
t.append(d[i]['solutions'][len(d[i]['solutions'])]['ms'])
#print len(t)
time = map(lambda x: x / 1000. / 60., t)
return time
listofalltimes = []
for fileitem in l:
d = zkl.load(fileitem)
alltimes = []
for u in range(UMAX):
partial_times = gettimes(d, u + 1)
alltimes.append(partial_times)
listofalltimes.append(alltimes)
#REFERENCE
#listofalltimes[0] = times for density 25
#listofalltimes[1] = times for density 30
#listofalltimes[2] = times for density 35
#listofalltimes[0][0] = times for density 25 with u = 2
#listofalltimes[0][1] = times for density 25 with u = 3
#listofalltimes[0][2] = times for density 25 with u = 4
for i in range(0,100):
if len(d[i]['solutions'])>=U:
if d[i]['solutions'][U]['eq'] == set([-1]): #supercliques
t.append(d[i]['solutions'][U-1]['ms'])
else:
t.append(d[i]['solutions'][U]['ms'])
else: #len of d[i]['solutions') < U
if d[i]['solutions'][len(d[i]['solutions'])]['eq'] == set([-1]): #supercliques
t.append(d[i]['solutions'][len(d[i]['solutions'])-1]['ms'])
else:
t.append(d[i]['solutions'][len(d[i]['solutions'])]['ms'])
time = map(lambda x: x/1000./60., t)
return time
d = zkl.load(l)
alltimes = []
mynames = []
for u in range(UMAX):
partial_times = gettimes(d,u+1)
alltimes.append(partial_times)
mynames.append(str(u+2))
g1 = sb.boxplot(alltimes,names=mynames,
widths=wds,fliersize=fliersz,
linewidth=lwd)
plt.xlabel('undersampling rate (u)')
plt.ylabel('computation time (minutes)')
plt.title('n = 5, number of graphs per u = 100, density = 25%' ,multialignment='center')
g1.figure.get_axes()[0].set_yscale('log')
def listplot(fname, mname='JJ', stl='', width=5, R=2):
l = zkl.load(fname)
l = l[:17 * 10]
y = min(width, len(l))
x = np.int(np.ceil(len(l) / float(y)))
d2l.matrix_list(l, y, x, R=R, w_gap=1, h_gap=2, mname=mname, stl=stl)
def timesfromfile(fname):
d = zkl.load(fname)
return getalltimes(d)