# if obj is not None:
# obj.train(20)
# f = open("som/output_si.txt",'w')
# positions = nx.get_node_attributes(obj.graph,'pos')
# for i,j in positions.iteritems():
# f.write(str(j)+"\n")
# f.close()
import fileinput
data, datamap = [], {}
for line in fileinput.input("som/output_si.txt"):
part = line.strip().replace("[","").replace("]","").split(",")
px, py = float(part[0])*5,1000-float(part[1])*5
data.append([px, py])
import pytesselate
import itertools
polygons = pytesselate.voronoi(data)
import pydraw
crs = pydraw.CoordinateSystem([0,0,1000,1000])
img = pydraw.Image(1000,1000)
for i,(center,poly) in enumerate(polygons):
print i, center
if i%8 == 0:
img.drawpolygon(poly,fillcolor=(0,0,0),outlinecolor=(0,0,0))
if i%8 == 1:
img.drawpolygon(poly,fillcolor=(255,0,0),outlinecolor=(0,0,0))
if i%8 == 2:
img.drawpolygon(poly,fillcolor=(0,255,0),outlinecolor=(0,0,0))
if i%8 == 3:
img.drawpolygon(poly,fillcolor=(0,0,255),outlinecolor=(0,0,0))
if i%8 == 4:
img.drawpolygon(poly,fillcolor=(0,255,255),outlinecolor=(0,0,0))
def func3():
# # draw grid
# GR = 50
# file = open("grid.txt","w")
# for i in xrange(GR+1):
# lat1, lat2 = 30.16, 30.34
# lng1 = 120.04+i*((120.24-120.04)/GR)
# lng2 = 120.04+i*((120.24-120.04)/GR)
# file.write("map.addOverlay(new BMap.Polyline([new BMap.Point("+str(lng1)+","+str(lat1)+"),new BMap.Point("+str(lng2)+","+str(lat2)+")], {strokeColor:\"red\", strokeWeight:1, strokeOpacity:0.25}));\n")
# for i in xrange(GR+1):
# lng1, lng2 = 120.04, 120.24
# lat1 = 30.16+i*((30.34-30.16)/GR)
# lat2 = 30.16+i*((30.34-30.16)/GR)
# file.write("map.addOverlay(new BMap.Polyline([new BMap.Point("+str(lng1)+", "+str(lat1)+"),new BMap.Point("+str(lng2)+", "+str(lat2)+")], {strokeColor:\"red\", strokeWeight:1, strokeOpacity:0.25}));\n")
# file.close()
# merge spark result
import fileinput
seen, num, sommap, assign = set(), 0, {}, {}
for line in fileinput.input("som/output_si.txt"):
part = line.strip().replace("[","").replace("]","").split(",")
if str(part[:2]) not in seen:
seen.add(str(part[:2]))
num, px, py = num+1, float(part[0]), float(part[1])
sommap[num] = {"center":[px,py],"wd_e":[0 for i in xrange(24)],"wd_d":[0 for i in xrange(24)],"we_e":[0 for i in xrange(24)],"we_d":[0 for i in xrange(24)]}
fileinput.close()
c = 0
for line in fileinput.input("data/movement/move_statistic.txt"):
c += 1
print c
part = line.strip().split("\t")[1:]
for p in part:
try:
a, w, i, s = p.split(" ")[0], int(p.split(" ")[1]), int(p.split(" ")[2]), int(p.split(" ")[3])
if not assign.has_key(tuple([int(a.split(",")[0]),int(a.split(",")[1])])):
sommin, distmin = 0, 0
for k,v in sommap.iteritems():
dist = euclidean(v["center"],[int(a.split(",")[0]),int(a.split(",")[1])])
if sommin == 0 or dist < distmin:
sommin, distmin = k, dist
assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])] = sommin
if w == 1 and s == 1:
sommap[assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])]]["wd_e"][i] += 1
if w == 1 and s == 0:
sommap[assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])]]["wd_d"][i] += 1
if w == 0 and s == 1:
sommap[assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])]]["we_e"][i] += 1
if w == 0 and s == 0:
sommap[assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])]]["we_d"][i] += 1
except:
continue
fileinput.close()
file = open("data/movement/move_statistic_merge.txt","w")
for k,v in sommap.iteritems():
file.write(str(k)+" "+" ".join([str(i) for i in v["wd_e"]])+" "+" ".join([str(i) for i in v["wd_d"]])+" "+" ".join([str(i) for i in v["we_e"]])+" "+" ".join([str(i) for i in v["we_d"]])+"\n")
file.close()
# generate document
file = open("GibbsLDA++/out.txt","w")
file.write("368\n")
c = 0
for line in fileinput.input("data/movement/move_statistic_merge.txt"):
part = line.split(" ")
grid, feature = part[0], [int(i) for i in part[1:]]
feature = [int(feature[i]/5.0) if 0<=i<24*2 else int(feature[i]/2.0) for i in xrange(len(feature))]
for i in xrange(24):
for j in xrange(feature[i]+1):
file.write(str(i)+",1,E ")
for i in xrange(24):
for j in xrange(feature[24+i]+1):
file.write(str(i)+",1,D ")
for i in xrange(24):
for j in xrange(feature[48+i]+1):
file.write(str(i)+",0,E ")
for i in xrange(24):
for j in xrange(feature[72+i]+1):
file.write(str(i)+",0,D ")
file.write("\n")
fileinput.close()
file.close()
# ./src/lda -est -alpha 0.5 -beta 0.5 -ntopics 8 -niters 1000 -savestep 2000 -twords 10 -dfile out.txt
# movement func 1
seen, somlist = set(), []
for line in fileinput.input("som/output_si.txt"):
part = line.strip().replace("[","").replace("]","").split(",")
if str(part[:2]) not in seen:
seen.add(str(part[:2]))
somlist.append([float(part[0])*5, 1000-float(part[1])*5, 0])
fileinput.close()
cnt = 0
for line in fileinput.input("GibbsLDA++/model_movement/model-final.theta"):
prob = [float(i) for i in line.strip().split(" ")]
somlist[cnt][2] = prob.index(max(prob))
cnt += 1
import pytesselate
import itertools
polygons = pytesselate.voronoi(somlist)
print len(polygons)
import pydraw
crs = pydraw.CoordinateSystem([0,0,1000,1000])
img = pydraw.Image(1000,1000)
for i,(center,poly) in enumerate(polygons):
if center:
if somlist[i][2] == 0:
img.drawpolygon(poly,fillcolor=(0,0,0),outlinecolor=(0,0,0))
if somlist[i][2] == 1:
img.drawpolygon(poly,fillcolor=(255,0,0),outlinecolor=(0,0,0))
if somlist[i][2] == 2:
img.drawpolygon(poly,fillcolor=(0,255,0),outlinecolor=(0,0,0))
if somlist[i][2] == 3:
img.drawpolygon(poly,fillcolor=(0,0,255),outlinecolor=(0,0,0))
if somlist[i][2] == 4:
img.drawpolygon(poly,fillcolor=(0,255,255),outlinecolor=(0,0,0))
if somlist[i][2] == 5:
img.drawpolygon(poly,fillcolor=(255,0,255),outlinecolor=(0,0,0))
if somlist[i][2] == 6:
img.drawpolygon(poly,fillcolor=(255,255,0),outlinecolor=(0,0,0))
if somlist[i][2] == 7:
img.drawpolygon(poly,fillcolor=(255,255,255),outlinecolor=(0,0,0))
# img.drawsquare(*center[:2], fillsize=2, fillcolor=(255,0,0))
img.save("test.png")
img.view()
# y0list, y1list, y2list, y3list = [], [], [], []
# # for line in fileinput.input("GibbsLDA++/model-final.phi"):
# for line in fileinput.input("LDA/model-final.phi"):
# prob = [float(i) for i in line.strip().split(" ")]
# y0, y1, y2, y3 = prob[0:24], prob[24:48], prob[48:72], prob[72:96]
# y0list.append(y0)
# y1list.append(y1)
# y2list.append(y2)
# y3list.append(y3)
# fig = plt.figure()
# ax1 = fig.add_subplot(241)
# ax1.plot([i for i in xrange(24)], y0list[0], c="red")
# ax1.plot([i for i in xrange(24)], y1list[0], c="yellow")
# ax1.plot([i for i in xrange(24)], y2list[0], c="green")
# ax1.plot([i for i in xrange(24)], y3list[0], c="blue")
# ax1.set_xlim(0,23)
# ax1.set_ylim(0,0.15)
# ax2 = fig.add_subplot(242)
# ax2.plot([i for i in xrange(24)], y0list[1], c="red")
# ax2.plot([i for i in xrange(24)], y1list[1], c="yellow")
# ax2.plot([i for i in xrange(24)], y2list[1], c="green")
# ax2.plot([i for i in xrange(24)], y3list[1], c="blue")
# ax2.set_xlim(0,23)
# ax2.set_ylim(0,0.15)
# ax3 = fig.add_subplot(243)
# ax3.plot([i for i in xrange(24)], y0list[2], c="red")
# ax3.plot([i for i in xrange(24)], y1list[2], c="yellow")
# ax3.plot([i for i in xrange(24)], y2list[2], c="green")
# ax3.plot([i for i in xrange(24)], y3list[2], c="blue")
# ax3.set_xlim(0,23)
# ax3.set_ylim(0,0.15)
# ax4 = fig.add_subplot(244)
# ax4.plot([i for i in xrange(24)], y0list[3], c="red")
# ax4.plot([i for i in xrange(24)], y1list[3], c="yellow")
# ax4.plot([i for i in xrange(24)], y2list[3], c="green")
# ax4.plot([i for i in xrange(24)], y3list[3], c="blue")
# ax4.set_xlim(0,23)
# ax4.set_ylim(0,0.15)
# ax5 = fig.add_subplot(245)
# ax5.plot([i for i in xrange(24)], y0list[4], c="red")
# ax5.plot([i for i in xrange(24)], y1list[4], c="yellow")
# ax5.plot([i for i in xrange(24)], y2list[4], c="green")
# ax5.plot([i for i in xrange(24)], y3list[4], c="blue")
# ax5.set_xlim(0,23)
# ax5.set_ylim(0,0.15)
# ax6 = fig.add_subplot(246)
# ax6.plot([i for i in xrange(24)], y0list[5], c="red")
# ax6.plot([i for i in xrange(24)], y1list[5], c="yellow")
# ax6.plot([i for i in xrange(24)], y2list[5], c="green")
# ax6.plot([i for i in xrange(24)], y3list[5], c="blue")
# ax6.set_xlim(0,23)
# ax6.set_ylim(0,0.15)
# ax7 = fig.add_subplot(247)
# ax7.plot([i for i in xrange(24)], y0list[6], c="red")
# ax7.plot([i for i in xrange(24)], y1list[6], c="yellow")
# ax7.plot([i for i in xrange(24)], y2list[6], c="green")
# ax7.plot([i for i in xrange(24)], y3list[6], c="blue")
# ax7.set_xlim(0,23)
# ax7.set_ylim(0,0.15)
# ax8 = fig.add_subplot(248)
# ax8.plot([i for i in xrange(24)], y0list[7], c="red")
# ax8.plot([i for i in xrange(24)], y1list[7], c="yellow")
# ax8.plot([i for i in xrange(24)], y2list[7], c="green")
# ax8.plot([i for i in xrange(24)], y3list[7], c="blue")
# ax8.set_xlim(0,23)
# ax8.set_ylim(0,0.15)
# plt.show()
# movement func 2
seen, somlist = set(), []
for line in fileinput.input("som/output_si.txt"):
part = line.strip().replace("[","").replace("]","").split(",")
if str(part[:2]) not in seen:
seen.add(str(part[:2]))
somlist.append([float(part[0])*5, 1000-float(part[1])*5, [0]*8])
fileinput.close()
print len(somlist)
cnt = 0
for line in fileinput.input("GibbsLDA++/model-final.theta"):
somlist[cnt][2] = [float(i) for i in line.strip().split(" ")]
cnt += 1
import pytesselate
import itertools
polygons = pytesselate.voronoi(somlist)
import pydraw
crs = pydraw.CoordinateSystem([0,0,1000,1000])
img = pydraw.Image(1000,1000)
c = 7
for i,(center,poly) in enumerate(polygons):
if center:
img.drawpolygon(poly,fillcolor=(255*somlist[i][2][c],255*somlist[i][2][c],255*somlist[i][2][c]),outlinecolor=(0,0,0))
img.save("figure/class_"+str(c)+".png")
# img.view()
# # poi func1
# tagmap = {}
# for line in fileinput.input("data/poi/poi.txt"):
# part = line.strip().split("\t")
# tag, lng, lat = part[0], float(part[1]), float(part[2])
# if 120.04<=lng<=120.24 and 30.16<=lat<=30.34:
# tagmap[tag] = 1 if not tagmap.has_key(tag) else tagmap[tag]+1
# fileinput.close()
# taglist = []
# for k,v in tagmap.iteritems():
# taglist.append({"k":k,"v":v})
# taglist = sorted(taglist, key=lambda x:x["v"])
# tag, count = [], []
# for item in taglist:
# if item["k"] in ["公司","购物","美食","房产","休闲","政府","丽人","金融","医疗","宾馆","汽车","教育","培训","旅游","运动"]:
# item["k"] = "company" if item["k"] == "公司"\
# else "shopping" if item["k"] == "购物"\
# else "food" if item["k"] == "美食"\
# else "housing" if item["k"] == "房产"\
# else "leisure" if item["k"] == "休闲"\
# else "government" if item["k"] == "政府"\
# else "beauty" if item["k"] == "丽人"\
# else "finance" if item["k"] == "金融"\
# else "medical" if item["k"] == "医疗"\
# else "hotel" if item["k"] == "宾馆"\
# else "car" if item["k"] == "汽车"\
# else "education" if item["k"] == "教育"\
# else "training" if item["k"] == "培训"\
# else "traveling" if item["k"] == "旅游"\
# else "physical" if item["k"] == "运动"\
# else ""
# tag.append(unicode(item["k"],"utf-8"))
# count.append(item["v"])
# for i in xrange(len(count)):
# print round(float(count[i])/sum(count),4)
# import matplotlib.pyplot as plt
# import numpy as np
# y_pos = np.arange(len(tag))
# plt.barh(y_pos, count, align='center', alpha=0.4)
# plt.yticks(y_pos, tag)
# plt.xlabel('count')
# plt.title('POI statistics')
# plt.show()
# # poi func2
# seen, num, sommap, assign = set(), 0, {}, {}
# for line in fileinput.input("som/output_si.txt"):
# part = line.strip().replace("[","").replace("]","").split(",")
# if str(part[:2]) not in seen:
# seen.add(str(part[:2]))
# px, py = float(part[0]), float(part[1])
# sommap[num] = {"center":[px,py],"poi":[0]*15}
# num += 1
# fileinput.close()
# for line in fileinput.input("data/poi/poi.txt"):
# part = line.strip().split("\t")
# tag, lng, lat = part[0], float(part[1]), float(part[2])
# if tag in ["公司","购物","美食","房产","休闲","政府","丽人","金融","医疗","宾馆","汽车","教育","培训","旅游","运动"] and 120.04<=lng<120.24 and 30.16<=lat<30.34:
# lng, lat = int((lng-120.04)/(120.24-120.04)*200), int((lat-30.16)/(30.34-30.16)*200)
# tag = 0 if tag == "公司"\
# else 1 if tag == "购物"\
# else 2 if tag == "美食"\
# else 3 if tag == "房产"\
# else 4 if tag == "休闲"\
# else 5 if tag == "政府"\
# else 6 if tag == "丽人"\
# else 7 if tag == "金融"\
# else 8 if tag == "医疗"\
# else 9 if tag == "宾馆"\
# else 10 if tag == "汽车"\
# else 11 if tag == "教育"\
# else 12 if tag == "培训"\
# else 13 if tag == "旅游"\
# else 14 if tag == "运动"\
# else -1
# if not assign.has_key(tuple([lng, lat])):
# sommin, distmin = 0, 0
# for k,v in sommap.iteritems():
# dist = euclidean(v["center"],[lng, lat])
# if sommin == 0 or dist < distmin:
# sommin, distmin = k, dist
# assign[tuple([lng, lat])] = sommin
# sommap[assign[tuple([lng, lat])]]["poi"][tag] += 1
# fileinput.close()
# classlist = [[0]*15 for i in xrange(8)]
# num = 0
# for line in fileinput.input("GibbsLDA++/model-final.theta"):
# prob = [float(i) for i in line.strip().split(" ")]
# for c in xrange(8):
# classlist[c] = [i+j for i,j in zip(classlist[c], [k*prob[c] for k in sommap[num]["poi"]])]
# num += 1
# fileinput.close()
# print classlist
# classlist = [[float(classlist[i][j])/sum(classlist[i]) for j in xrange(15)] for i in xrange(8)]
# avg = [0.2304,0.2043,0.1472,0.1332,0.0504,0.0393,0.0391,0.0321,0.0263,0.0242,0.0223,0.0185,0.0128,0.0100,0.0099]
# classlist = [[classlist[i][j]-avg[j] for j in xrange(15)] for i in xrange(8)]
# import matplotlib.pyplot as plt
# import numpy as np
# fig = plt.figure()
# ax1 = fig.add_subplot(241)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[0], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax1.set_ylim(-0.1,0.1)
# ax2 = fig.add_subplot(242)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[1], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax2.set_ylim(-0.1,0.1)
# ax3 = fig.add_subplot(243)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[2], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax3.set_ylim(-0.1,0.1)
# ax4 = fig.add_subplot(244)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[3], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax4.set_ylim(-0.1,0.1)
# ax5 = fig.add_subplot(245)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[4], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax5.set_ylim(-0.1,0.1)
# ax6 = fig.add_subplot(246)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[5], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax6.set_ylim(-0.1,0.1)
# ax7 = fig.add_subplot(247)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[6], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax7.set_ylim(-0.1,0.1)
# ax8 = fig.add_subplot(248)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[7], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax8.set_ylim(-0.1,0.1)
# plt.show()
# generate document poi
seen, num, sommap, assign = set(), 0, {}, {}
for line in fileinput.input("som/output_si.txt"):
part = line.strip().replace("[","").replace("]","").split(",")
if str(part[:2]) not in seen:
seen.add(str(part[:2]))
px, py = float(part[0]), float(part[1])
sommap[num] = {"center":[px,py],"poi":[0]*15}
num += 1
fileinput.close()
count = 0
for line in fileinput.input("data/poi/poi.txt"):
count += 1
print count
part = line.strip().split("\t")
tag, lng, lat = part[0], float(part[1]), float(part[2])
if tag in ["公司","购物","美食","房产","休闲","政府","丽人","金融","医疗","宾馆","汽车","教育","培训","旅游","运动"] and 120.04<=lng<120.24 and 30.16<=lat<30.34:
lng, lat = int((lng-120.04)/(120.24-120.04)*200), int((lat-30.16)/(30.34-30.16)*200)
tag = 0 if tag == "公司"\
else 1 if tag == "购物"\
else 2 if tag == "美食"\
else 3 if tag == "房产"\
else 4 if tag == "休闲"\
else 5 if tag == "政府"\
else 6 if tag == "丽人"\
else 7 if tag == "金融"\
else 8 if tag == "医疗"\
else 9 if tag == "宾馆"\
else 10 if tag == "汽车"\
else 11 if tag == "教育"\
else 12 if tag == "培训"\
else 13 if tag == "旅游"\
else 14 if tag == "运动"\
else -1
if not assign.has_key(tuple([lng, lat])):
sommin, distmin = 0, 0
for k,v in sommap.iteritems():
dist = euclidean(v["center"],[lng, lat])
if sommin == 0 or dist < distmin:
sommin, distmin = k, dist
assign[tuple([lng, lat])] = sommin
sommap[assign[tuple([lng, lat])]]["poi"][tag] += 1
fileinput.close()
file = open("GibbsLDA++/out_poi.txt","w")
file.write("368\n")
for k,v in sommap.iteritems():
file.write("-1 "+" ".join([" ".join([str(x) for y in xrange(v["poi"][x])]) for x in xrange(15)])+"\n")
file.close()
# ./src/lda -est -alpha 0.5 -beta 0.5 -ntopics 8 -niters 1000 -savestep 2000 -twords 10 -dfile out_poi.txt
# movement func 1
seen, somlist = set(), []
for line in fileinput.input("som/output_si.txt"):
part = line.strip().replace("[","").replace("]","").split(",")
if str(part[:2]) not in seen:
seen.add(str(part[:2]))
somlist.append([float(part[0])*5, 1000-float(part[1])*5, 0])
fileinput.close()
cnt = 0
for line in fileinput.input("GibbsLDA++/model-final.theta"):
prob = [float(i) for i in line.strip().split(" ")]
somlist[cnt][2] = prob.index(max(prob))
cnt += 1
import pytesselate
import itertools
polygons = pytesselate.voronoi(somlist)
print len(polygons)
import pydraw
crs = pydraw.CoordinateSystem([0,0,1000,1000])
img = pydraw.Image(1000,1000)
for i,(center,poly) in enumerate(polygons):
if center:
if somlist[i][2] == 0:
img.drawpolygon(poly,fillcolor=(0,0,0),outlinecolor=(0,0,0))
if somlist[i][2] == 1:
img.drawpolygon(poly,fillcolor=(255,0,0),outlinecolor=(0,0,0))
if somlist[i][2] == 2:
img.drawpolygon(poly,fillcolor=(0,255,0),outlinecolor=(0,0,0))
if somlist[i][2] == 3:
img.drawpolygon(poly,fillcolor=(0,0,255),outlinecolor=(0,0,0))
if somlist[i][2] == 4:
img.drawpolygon(poly,fillcolor=(0,255,255),outlinecolor=(0,0,0))
if somlist[i][2] == 5:
img.drawpolygon(poly,fillcolor=(255,0,255),outlinecolor=(0,0,0))
if somlist[i][2] == 6:
img.drawpolygon(poly,fillcolor=(255,255,0),outlinecolor=(0,0,0))
if somlist[i][2] == 7:
img.drawpolygon(poly,fillcolor=(255,255,255),outlinecolor=(0,0,0))
# img.drawsquare(*center[:2], fillsize=2, fillcolor=(255,0,0))
img.save("test.png")
img.view()
# ylist = []
# for line in fileinput.input("GibbsLDA++/model-final.phi"):
# prob = [float(i) for i in line.strip().split(" ")]
# ylist.append(prob)
# print len(ylist[0])
# fig = plt.figure()
# ax1 = fig.add_subplot(241)
# ax1.plot([i for i in xrange(16)], ylist[0], c="red")
# ax1.set_xlim(0,15)
# ax1.set_ylim(0,1.0)
# ax2 = fig.add_subplot(242)
# ax2.plot([i for i in xrange(16)], ylist[1], c="red")
# ax2.set_xlim(0,15)
# ax2.set_ylim(0,1.0)
# ax3 = fig.add_subplot(243)
# ax3.plot([i for i in xrange(16)], ylist[2], c="red")
# ax3.set_xlim(0,15)
# ax3.set_ylim(0,1.0)
# ax4 = fig.add_subplot(244)
# ax4.plot([i for i in xrange(16)], ylist[3], c="red")
# ax4.set_xlim(0,15)
# ax4.set_ylim(0,1.0)
# ax5 = fig.add_subplot(245)
# ax5.plot([i for i in xrange(16)], ylist[4], c="red")
# ax5.set_xlim(0,15)
# ax5.set_ylim(0,1.0)
# ax6 = fig.add_subplot(246)
# ax6.plot([i for i in xrange(16)], ylist[5], c="red")
# ax6.set_xlim(0,15)
# ax6.set_ylim(0,1.0)
# ax7 = fig.add_subplot(247)
# ax7.plot([i for i in xrange(16)], ylist[6], c="red")
# ax7.set_xlim(0,15)
# ax7.set_ylim(0,1.0)
# ax8 = fig.add_subplot(248)
# ax8.plot([i for i in xrange(16)], ylist[7], c="red")
# ax8.set_xlim(0,15)
# ax8.set_ylim(0,1.0)
# plt.show()
# movement func 2
seen, somlist = set(), []
for line in fileinput.input("som/output_si.txt"):
part = line.strip().replace("[","").replace("]","").split(",")
if str(part[:2]) not in seen:
seen.add(str(part[:2]))
somlist.append([float(part[0])*5, 1000-float(part[1])*5, [0]*8])
fileinput.close()
print len(somlist)
cnt = 0
for line in fileinput.input("GibbsLDA++/model-final.theta"):
somlist[cnt][2] = [float(i) for i in line.strip().split(" ")]
cnt += 1
import pytesselate
import itertools
polygons = pytesselate.voronoi(somlist)
import pydraw
crs = pydraw.CoordinateSystem([0,0,1000,1000])
img = pydraw.Image(1000,1000)
c = 6
for i,(center,poly) in enumerate(polygons):
if center:
img.drawpolygon(poly,fillcolor=(255*somlist[i][2][c],255*somlist[i][2][c],255*somlist[i][2][c]),outlinecolor=(0,0,0))
img.save("figure/class_"+str(c)+".png")
def func4():
# # draw grid
# GR = 50
# file = open("grid.txt","w")
# for i in xrange(GR+1):
# lat1, lat2 = 30.16, 30.34
# lng1 = 120.04+i*((120.24-120.04)/GR)
# lng2 = 120.04+i*((120.24-120.04)/GR)
# file.write("map.addOverlay(new BMap.Polyline([new BMap.Point("+str(lng1)+","+str(lat1)+"),new BMap.Point("+str(lng2)+","+str(lat2)+")], {strokeColor:\"red\", strokeWeight:1, strokeOpacity:0.25}));\n")
# for i in xrange(GR+1):
# lng1, lng2 = 120.04, 120.24
# lat1 = 30.16+i*((30.34-30.16)/GR)
# lat2 = 30.16+i*((30.34-30.16)/GR)
# file.write("map.addOverlay(new BMap.Polyline([new BMap.Point("+str(lng1)+", "+str(lat1)+"),new BMap.Point("+str(lng2)+", "+str(lat2)+")], {strokeColor:\"red\", strokeWeight:1, strokeOpacity:0.25}));\n")
# file.close()
# # merge spark result
# import fileinput
# seen, num, sommap, assign = set(), 0, {}, {}
# for line in fileinput.input("som/output_si.txt"):
# part = line.strip().replace("[","").replace("]","").split(",")
# if str(part[:2]) not in seen:
# seen.add(str(part[:2]))
# num, px, py = num+1, float(part[0]), float(part[1])
# sommap[num] = {"center":[px,py],"wd_e":[0 for i in xrange(24)],"wd_d":[0 for i in xrange(24)],"we_e":[0 for i in xrange(24)],"we_d":[0 for i in xrange(24)]}
# fileinput.close()
# c = 0
# for line in fileinput.input("data/movement/move_statistic.txt"):
# c += 1
# print c
# part = line.strip().split("\t")[1:]
# for p in part:
# try:
# a, w, i, s = p.split(" ")[0], int(p.split(" ")[1]), int(p.split(" ")[2]), int(p.split(" ")[3])
# if not assign.has_key(tuple([int(a.split(",")[0]),int(a.split(",")[1])])):
# sommin, distmin = 0, 0
# for k,v in sommap.iteritems():
# dist = euclidean(v["center"],[int(a.split(",")[0]),int(a.split(",")[1])])
# if sommin == 0 or dist < distmin:
# sommin, distmin = k, dist
# assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])] = sommin
# if w == 1 and s == 1:
# sommap[assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])]]["wd_e"][i] += 1
# if w == 1 and s == 0:
# sommap[assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])]]["wd_d"][i] += 1
# if w == 0 and s == 1:
# sommap[assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])]]["we_e"][i] += 1
# if w == 0 and s == 0:
# sommap[assign[tuple([int(a.split(",")[0]),int(a.split(",")[1])])]]["we_d"][i] += 1
# except:
# continue
# fileinput.close()
# file = open("data/movement/move_statistic_merge.txt","w")
# for k,v in sommap.iteritems():
# file.write(str(k)+" "+" ".join([str(i) for i in v["wd_e"]])+" "+" ".join([str(i) for i in v["wd_d"]])+" "+" ".join([str(i) for i in v["we_e"]])+" "+" ".join([str(i) for i in v["we_d"]])+"\n")
# file.close()
# # generate document
# file = open("GibbsLDA++/out.txt","w")
# file.write("368\n")
# c = 0
# for line in fileinput.input("data/movement/move_statistic_merge.txt"):
# part = line.split(" ")
# grid, feature = part[0], [int(i) for i in part[1:]]
# feature = [int(feature[i]/5.0) if 0<=i<24*2 else int(feature[i]/2.0) for i in xrange(len(feature))]
# for i in xrange(24):
# for j in xrange(feature[i]+1):
# file.write(str(i)+",1,E ")
# for i in xrange(24):
# for j in xrange(feature[24+i]+1):
# file.write(str(i)+",1,D ")
# for i in xrange(24):
# for j in xrange(feature[48+i]+1):
# file.write(str(i)+",0,E ")
# for i in xrange(24):
# for j in xrange(feature[72+i]+1):
# file.write(str(i)+",0,D ")
# file.write("\n")
# fileinput.close()
# file.close()
# ./src/lda -est -alpha 0.5 -beta 0.5 -ntopics 8 -niters 1000 -savestep 2000 -twords 10 -dfile out.txt
# movement func 1
seen, somlist = set(), []
for line in fileinput.input("som/output_si.txt"):
part = line.strip().replace("[","").replace("]","").split(",")
if str(part[:2]) not in seen:
seen.add(str(part[:2]))
somlist.append([float(part[0])*5, 1000-float(part[1])*5, 0])
fileinput.close()
cnt = 0
# for line in fileinput.input("GibbsLDA++/model-final.theta"):
for line in fileinput.input("LDA/model-final.theta"):
prob = [float(i) for i in line.strip().split(" ")]
somlist[cnt][2] = prob.index(max(prob))
cnt += 1
import pytesselate
import itertools
polygons = pytesselate.voronoi(somlist)
print len(polygons)
import pydraw
crs = pydraw.CoordinateSystem([0,0,1000,1000])
img = pydraw.Image(1000,1000)
for i,(center,poly) in enumerate(polygons):
if center:
if somlist[i][2] == 0:
img.drawpolygon(poly,fillcolor=(0,0,0),outlinecolor=(0,0,0))
if somlist[i][2] == 1:
img.drawpolygon(poly,fillcolor=(255,0,0),outlinecolor=(0,0,0))
if somlist[i][2] == 2:
img.drawpolygon(poly,fillcolor=(0,255,0),outlinecolor=(0,0,0))
if somlist[i][2] == 3:
img.drawpolygon(poly,fillcolor=(0,0,255),outlinecolor=(0,0,0))
if somlist[i][2] == 4:
img.drawpolygon(poly,fillcolor=(0,255,255),outlinecolor=(0,0,0))
if somlist[i][2] == 5:
img.drawpolygon(poly,fillcolor=(255,0,255),outlinecolor=(0,0,0))
if somlist[i][2] == 6:
img.drawpolygon(poly,fillcolor=(255,255,0),outlinecolor=(0,0,0))
if somlist[i][2] == 7:
img.drawpolygon(poly,fillcolor=(255,255,255),outlinecolor=(0,0,0))
# img.drawsquare(*center[:2], fillsize=2, fillcolor=(255,0,0))
img.save("test.png")
img.view()
y0list, y1list, y2list, y3list = [], [], [], []
# for line in fileinput.input("GibbsLDA++/model-final.phi"):
for line in fileinput.input("LDA/model-final.phi"):
prob = [float(i) for i in line.strip().split(" ")]
y0, y1, y2, y3 = prob[0:24], prob[24:48], prob[48:72], prob[72:96]
y0list.append(y0)
y1list.append(y1)
y2list.append(y2)
y3list.append(y3)
fig = plt.figure()
ax1 = fig.add_subplot(241)
ax1.plot([i for i in xrange(24)], y0list[0], c="red")
ax1.plot([i for i in xrange(24)], y1list[0], c="yellow")
ax1.plot([i for i in xrange(24)], y2list[0], c="green")
ax1.plot([i for i in xrange(24)], y3list[0], c="blue")
ax1.set_xlim(0,23)
ax1.set_ylim(0,0.15)
ax2 = fig.add_subplot(242)
ax2.plot([i for i in xrange(24)], y0list[1], c="red")
ax2.plot([i for i in xrange(24)], y1list[1], c="yellow")
ax2.plot([i for i in xrange(24)], y2list[1], c="green")
ax2.plot([i for i in xrange(24)], y3list[1], c="blue")
ax2.set_xlim(0,23)
ax2.set_ylim(0,0.15)
ax3 = fig.add_subplot(243)
ax3.plot([i for i in xrange(24)], y0list[2], c="red")
ax3.plot([i for i in xrange(24)], y1list[2], c="yellow")
ax3.plot([i for i in xrange(24)], y2list[2], c="green")
ax3.plot([i for i in xrange(24)], y3list[2], c="blue")
ax3.set_xlim(0,23)
ax3.set_ylim(0,0.15)
ax4 = fig.add_subplot(244)
ax4.plot([i for i in xrange(24)], y0list[3], c="red")
ax4.plot([i for i in xrange(24)], y1list[3], c="yellow")
ax4.plot([i for i in xrange(24)], y2list[3], c="green")
ax4.plot([i for i in xrange(24)], y3list[3], c="blue")
ax4.set_xlim(0,23)
ax4.set_ylim(0,0.15)
ax5 = fig.add_subplot(245)
ax5.plot([i for i in xrange(24)], y0list[4], c="red")
ax5.plot([i for i in xrange(24)], y1list[4], c="yellow")
ax5.plot([i for i in xrange(24)], y2list[4], c="green")
ax5.plot([i for i in xrange(24)], y3list[4], c="blue")
ax5.set_xlim(0,23)
ax5.set_ylim(0,0.15)
ax6 = fig.add_subplot(246)
ax6.plot([i for i in xrange(24)], y0list[5], c="red")
ax6.plot([i for i in xrange(24)], y1list[5], c="yellow")
ax6.plot([i for i in xrange(24)], y2list[5], c="green")
ax6.plot([i for i in xrange(24)], y3list[5], c="blue")
ax6.set_xlim(0,23)
ax6.set_ylim(0,0.15)
ax7 = fig.add_subplot(247)
ax7.plot([i for i in xrange(24)], y0list[6], c="red")
ax7.plot([i for i in xrange(24)], y1list[6], c="yellow")
ax7.plot([i for i in xrange(24)], y2list[6], c="green")
ax7.plot([i for i in xrange(24)], y3list[6], c="blue")
ax7.set_xlim(0,23)
ax7.set_ylim(0,0.15)
ax8 = fig.add_subplot(248)
ax8.plot([i for i in xrange(24)], y0list[7], c="red")
ax8.plot([i for i in xrange(24)], y1list[7], c="yellow")
ax8.plot([i for i in xrange(24)], y2list[7], c="green")
ax8.plot([i for i in xrange(24)], y3list[7], c="blue")
ax8.set_xlim(0,23)
ax8.set_ylim(0,0.15)
plt.show()
# # movement func 2
# seen, somlist = set(), []
# for line in fileinput.input("som/output_si.txt"):
# part = line.strip().replace("[","").replace("]","").split(",")
# if str(part[:2]) not in seen:
# seen.add(str(part[:2]))
# somlist.append([float(part[0])*5, 1000-float(part[1])*5, [0]*8])
# fileinput.close()
# print len(somlist)
# cnt = 0
# for line in fileinput.input("GibbsLDA++/model-final.theta"):
# somlist[cnt][2] = [float(i) for i in line.strip().split(" ")]
# cnt += 1
# import pytesselate
# import itertools
# polygons = pytesselate.voronoi(somlist)
# import pydraw
# crs = pydraw.CoordinateSystem([0,0,1000,1000])
# img = pydraw.Image(1000,1000)
# c = 7
# for i,(center,poly) in enumerate(polygons):
# if center:
# img.drawpolygon(poly,fillcolor=(255*somlist[i][2][c],255*somlist[i][2][c],255*somlist[i][2][c]),outlinecolor=(0,0,0))
# img.save("figure/class_"+str(c)+".png")
# # img.view()
# # poi func1
# tagmap = {}
# for line in fileinput.input("data/poi/poi.txt"):
# part = line.strip().split("\t")
# tag, lng, lat = part[0], float(part[1]), float(part[2])
# if 120.04<=lng<=120.24 and 30.16<=lat<=30.34:
# tagmap[tag] = 1 if not tagmap.has_key(tag) else tagmap[tag]+1
# fileinput.close()
# taglist = []
# for k,v in tagmap.iteritems():
# taglist.append({"k":k,"v":v})
# taglist = sorted(taglist, key=lambda x:x["v"])
# tag, count = [], []
# for item in taglist:
# if item["k"] in ["公司","购物","美食","房产","休闲","政府","丽人","金融","医疗","宾馆","汽车","教育","培训","旅游","运动"]:
# item["k"] = "company" if item["k"] == "公司"\
# else "shopping" if item["k"] == "购物"\
# else "food" if item["k"] == "美食"\
# else "housing" if item["k"] == "房产"\
# else "leisure" if item["k"] == "休闲"\
# else "government" if item["k"] == "政府"\
# else "beauty" if item["k"] == "丽人"\
# else "finance" if item["k"] == "金融"\
# else "medical" if item["k"] == "医疗"\
# else "hotel" if item["k"] == "宾馆"\
# else "car" if item["k"] == "汽车"\
# else "education" if item["k"] == "教育"\
# else "training" if item["k"] == "培训"\
# else "traveling" if item["k"] == "旅游"\
# else "physical" if item["k"] == "运动"\
# else ""
# tag.append(unicode(item["k"],"utf-8"))
# count.append(item["v"])
# for i in xrange(len(count)):
# print round(float(count[i])/sum(count),4)
# import matplotlib.pyplot as plt
# import numpy as np
# y_pos = np.arange(len(tag))
# plt.barh(y_pos, count, align='center', alpha=0.4)
# plt.yticks(y_pos, tag)
# plt.xlabel('count')
# plt.title('POI statistics')
# plt.show()
# # poi func2
# seen, num, sommap, assign = set(), 0, {}, {}
# for line in fileinput.input("som/output_si.txt"):
# part = line.strip().replace("[","").replace("]","").split(",")
# if str(part[:2]) not in seen:
# seen.add(str(part[:2]))
# px, py = float(part[0]), float(part[1])
# sommap[num] = {"center":[px,py],"poi":[0]*15}
# num += 1
# fileinput.close()
# for line in fileinput.input("data/poi/poi.txt"):
# part = line.strip().split("\t")
# tag, lng, lat = part[0], float(part[1]), float(part[2])
# if tag in ["公司","购物","美食","房产","休闲","政府","丽人","金融","医疗","宾馆","汽车","教育","培训","旅游","运动"] and 120.04<=lng<120.24 and 30.16<=lat<30.34:
# lng, lat = int((lng-120.04)/(120.24-120.04)*200), int((lat-30.16)/(30.34-30.16)*200)
# tag = 0 if tag == "公司"\
# else 1 if tag == "购物"\
# else 2 if tag == "美食"\
# else 3 if tag == "房产"\
# else 4 if tag == "休闲"\
# else 5 if tag == "政府"\
# else 6 if tag == "丽人"\
# else 7 if tag == "金融"\
# else 8 if tag == "医疗"\
# else 9 if tag == "宾馆"\
# else 10 if tag == "汽车"\
# else 11 if tag == "教育"\
# else 12 if tag == "培训"\
# else 13 if tag == "旅游"\
# else 14 if tag == "运动"\
# else -1
# if not assign.has_key(tuple([lng, lat])):
# sommin, distmin = 0, 0
# for k,v in sommap.iteritems():
# dist = euclidean(v["center"],[lng, lat])
# if sommin == 0 or dist < distmin:
# sommin, distmin = k, dist
# assign[tuple([lng, lat])] = sommin
# sommap[assign[tuple([lng, lat])]]["poi"][tag] += 1
# fileinput.close()
# classlist = [[0]*15 for i in xrange(8)]
# num = 0
# for line in fileinput.input("GibbsLDA++/model-final.theta"):
# prob = [float(i) for i in line.strip().split(" ")]
# for c in xrange(8):
# classlist[c] = [i+j for i,j in zip(classlist[c], [k*prob[c] for k in sommap[num]["poi"]])]
# num += 1
# fileinput.close()
# print classlist
# classlist = [[float(classlist[i][j])/sum(classlist[i]) for j in xrange(15)] for i in xrange(8)]
# avg = [0.2304,0.2043,0.1472,0.1332,0.0504,0.0393,0.0391,0.0321,0.0263,0.0242,0.0223,0.0185,0.0128,0.0100,0.0099]
# classlist = [[classlist[i][j]-avg[j] for j in xrange(15)] for i in xrange(8)]
# import matplotlib.pyplot as plt
# import numpy as np
# fig = plt.figure()
# ax1 = fig.add_subplot(241)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[0], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax1.set_ylim(-0.1,0.1)
# ax2 = fig.add_subplot(242)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[1], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax2.set_ylim(-0.1,0.1)
# ax3 = fig.add_subplot(243)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[2], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax3.set_ylim(-0.1,0.1)
# ax4 = fig.add_subplot(244)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[3], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax4.set_ylim(-0.1,0.1)
# ax5 = fig.add_subplot(245)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[4], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax5.set_ylim(-0.1,0.1)
# ax6 = fig.add_subplot(246)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[5], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax6.set_ylim(-0.1,0.1)
# ax7 = fig.add_subplot(247)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[6], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax7.set_ylim(-0.1,0.1)
# ax8 = fig.add_subplot(248)
# tag = [str(i) for i in xrange(15)]
# y_pos = np.arange(len(tag))
# plt.bar(y_pos, classlist[7], align='center', alpha=0.4)
# plt.xticks(y_pos, tag)
# ax8.set_ylim(-0.1,0.1)
# plt.show()
# # generate document poi
# seen, num, sommap, assign = set(), 0, {}, {}
# for line in fileinput.input("som/output_si.txt"):
# part = line.strip().replace("[","").replace("]","").split(",")
# if str(part[:2]) not in seen:
# seen.add(str(part[:2]))
# px, py = float(part[0]), float(part[1])
# sommap[num] = {"center":[px,py],"poi":[0]*15}
# num += 1
# fileinput.close()
# count = 0
# for line in fileinput.input("data/poi/poi.txt"):
# count += 1
# print count
# part = line.strip().split("\t")
# tag, lng, lat = part[0], float(part[1]), float(part[2])
# if tag in ["公司","购物","美食","房产","休闲","政府","丽人","金融","医疗","宾馆","汽车","教育","培训","旅游","运动"] and 120.04<=lng<120.24 and 30.16<=lat<30.34:
# lng, lat = int((lng-120.04)/(120.24-120.04)*200), int((lat-30.16)/(30.34-30.16)*200)
# tag = 0 if tag == "公司"\
# else 1 if tag == "购物"\
# else 2 if tag == "美食"\
# else 3 if tag == "房产"\
# else 4 if tag == "休闲"\
# else 5 if tag == "政府"\
# else 6 if tag == "丽人"\
# else 7 if tag == "金融"\
# else 8 if tag == "医疗"\
# else 9 if tag == "宾馆"\
# else 10 if tag == "汽车"\
# else 11 if tag == "教育"\
# else 12 if tag == "培训"\
# else 13 if tag == "旅游"\
# else 14 if tag == "运动"\
# else -1
# if not assign.has_key(tuple([lng, lat])):
# sommin, distmin = 0, 0
# for k,v in sommap.iteritems():
# dist = euclidean(v["center"],[lng, lat])
# if sommin == 0 or dist < distmin:
# sommin, distmin = k, dist
# assign[tuple([lng, lat])] = sommin
# sommap[assign[tuple([lng, lat])]]["poi"][tag] += 1
# fileinput.close()
# file = open("GibbsLDA++/out_poi.txt","w")
# file.write("368\n")
# for k,v in sommap.iteritems():
# file.write("-1 "+" ".join([" ".join([str(x) for y in xrange(v["poi"][x])]) for x in xrange(15)])+"\n")
# file.close()
# ./src/lda -est -alpha 0.5 -beta 0.5 -ntopics 8 -niters 1000 -savestep 2000 -twords 10 -dfile out_poi.txt
# # movement func 1
# seen, somlist = set(), []
# for line in fileinput.input("som/output_si.txt"):
# part = line.strip().replace("[","").replace("]","").split(",")
# if str(part[:2]) not in seen:
# seen.add(str(part[:2]))
# somlist.append([float(part[0])*5, 1000-float(part[1])*5, 0])
# fileinput.close()
# cnt = 0
# for line in fileinput.input("GibbsLDA++/model-final.theta"):
# prob = [float(i) for i in line.strip().split(" ")]
# somlist[cnt][2] = prob.index(max(prob))
# cnt += 1
# import pytesselate
# import itertools
# polygons = pytesselate.voronoi(somlist)
# print len(polygons)
# import pydraw
# crs = pydraw.CoordinateSystem([0,0,1000,1000])
# img = pydraw.Image(1000,1000)
# for i,(center,poly) in enumerate(polygons):
# if center:
# if somlist[i][2] == 0:
# img.drawpolygon(poly,fillcolor=(0,0,0),outlinecolor=(0,0,0))
# if somlist[i][2] == 1:
# img.drawpolygon(poly,fillcolor=(255,0,0),outlinecolor=(0,0,0))
# if somlist[i][2] == 2:
# img.drawpolygon(poly,fillcolor=(0,255,0),outlinecolor=(0,0,0))
# if somlist[i][2] == 3:
# img.drawpolygon(poly,fillcolor=(0,0,255),outlinecolor=(0,0,0))
# if somlist[i][2] == 4:
# img.drawpolygon(poly,fillcolor=(0,255,255),outlinecolor=(0,0,0))
# if somlist[i][2] == 5:
# img.drawpolygon(poly,fillcolor=(255,0,255),outlinecolor=(0,0,0))
# if somlist[i][2] == 6:
# img.drawpolygon(poly,fillcolor=(255,255,0),outlinecolor=(0,0,0))
# if somlist[i][2] == 7:
# img.drawpolygon(poly,fillcolor=(255,255,255),outlinecolor=(0,0,0))
# # img.drawsquare(*center[:2], fillsize=2, fillcolor=(255,0,0))
# img.save("test.png")
# img.view()
# ylist = []
# for line in fileinput.input("GibbsLDA++/model-final.phi"):
# prob = [float(i) for i in line.strip().split(" ")]
# ylist.append(prob)
# print len(ylist[0])
# fig = plt.figure()
# ax1 = fig.add_subplot(241)
# ax1.plot([i for i in xrange(16)], ylist[0], c="red")
# ax1.set_xlim(0,15)
# ax1.set_ylim(0,1.0)
# ax2 = fig.add_subplot(242)
# ax2.plot([i for i in xrange(16)], ylist[1], c="red")
# ax2.set_xlim(0,15)
# ax2.set_ylim(0,1.0)
# ax3 = fig.add_subplot(243)
# ax3.plot([i for i in xrange(16)], ylist[2], c="red")
# ax3.set_xlim(0,15)
# ax3.set_ylim(0,1.0)
# ax4 = fig.add_subplot(244)
# ax4.plot([i for i in xrange(16)], ylist[3], c="red")
# ax4.set_xlim(0,15)
# ax4.set_ylim(0,1.0)
# ax5 = fig.add_subplot(245)
# ax5.plot([i for i in xrange(16)], ylist[4], c="red")
# ax5.set_xlim(0,15)
# ax5.set_ylim(0,1.0)
# ax6 = fig.add_subplot(246)
# ax6.plot([i for i in xrange(16)], ylist[5], c="red")
# ax6.set_xlim(0,15)
# ax6.set_ylim(0,1.0)
# ax7 = fig.add_subplot(247)
# ax7.plot([i for i in xrange(16)], ylist[6], c="red")
# ax7.set_xlim(0,15)
# ax7.set_ylim(0,1.0)
# ax8 = fig.add_subplot(248)
# ax8.plot([i for i in xrange(16)], ylist[7], c="red")
# ax8.set_xlim(0,15)
# ax8.set_ylim(0,1.0)
# plt.show()
# # movement func 2
# seen, somlist = set(), []
# for line in fileinput.input("som/output_si.txt"):
# part = line.strip().replace("[","").replace("]","").split(",")
# if str(part[:2]) not in seen:
# seen.add(str(part[:2]))
# somlist.append([float(part[0])*5, 1000-float(part[1])*5, [0]*8])
# fileinput.close()
# print len(somlist)
# cnt = 0
# for line in fileinput.input("GibbsLDA++/model-final.theta"):
# somlist[cnt][2] = [float(i) for i in line.strip().split(" ")]
# cnt += 1
# import pytesselate
# import itertools
# polygons = pytesselate.voronoi(somlist)
# import pydraw
# crs = pydraw.CoordinateSystem([0,0,1000,1000])
# img = pydraw.Image(1000,1000)
# c = 6
# for i,(center,poly) in enumerate(polygons):
# if center:
# img.drawpolygon(poly,fillcolor=(255*somlist[i][2][c],255*somlist[i][2][c],255*somlist[i][2][c]),outlinecolor=(0,0,0))
# img.save("figure/class_"+str(c)+".png")
# # img.view()
pass
