myprojects-hg-reborn

Fork

#!/usr/bin/env python
import sys, time
import os
from numpy import *
import sociopatterns
from sociopatterns.loader import Loader
from sociopatterns.analysis import ContactGraphAnalyzer
from xml.dom import minidom
import string
import scipy as s

from sociopatterns.loader import TagMapping

M = TagMapping()

f = open('tag_mapping_high_presence.dat')
for line in f:
    (mapped_id, orig_id, t1, t2, type) = string.split(line[:-1])
    mapped_id = int(mapped_id)
    orig_id = int(orig_id)
    t1 = int(t1)
    t2 = int(t2)
    M.add_mapping(orig_id, mapped_id, t1, t2)
f.close()


# ================= CONFIGURE THESE PARAMETERS ==========================
XXTEA_CRYPTO_KEY = ( 0xf6e103d4, 0x77a739f6, 0x65eecead, 0xa40543a9 )
DECODE = 0
EXPERIMENT = "OBG"
TAG_ID_MIN = 10
TAG_ID_MAX = +3000
DELTAT = 120         # aggregation window
K_MAX = 24           # filter out all nodes with degree > K_MAX (useful for filtering away clusters of tags left together)
# =======================================================================

TAG_ID_SPAN = TAG_ID_MAX - TAG_ID_MIN

cum_adj = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
onoff = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
adj_prev = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
time_first_seen = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
time_last_seen = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
frame_count = 0




####################################################################################





def collect_node_and_edge_properties(TAG_ID_SPAN,direct_indirect_seen,TAG_ID_MIN,cum_adj,onoff,time_first_seen, time_last_seen):

    node_attr_list=[]


    node_label_list=[]

    edge_attr_list=[]

    edge_label_list=[]


    for i in range(TAG_ID_SPAN):
        if not direct_indirect_seen[i]:
            continue

        node_attr_list.append("[")

        node_attr_list.append('%d' % i)
        node_attr_list.append('"%d"' % (TAG_ID_MIN+i))



        node_label_list.append("node")
        node_label_list.append("id")
        node_label_list.append("label")




        if cum_adj[i, i]:

            node_attr_list.append('%d' % cum_adj[i,i])
            node_attr_list.append('%d' % onoff[i,i])
            node_attr_list.append('%e' %  (onoff[i,i] / float(cum_adj[i,i]) / 2.0) )
            node_attr_list.append('%d' % time_first_seen[i,i])
            node_attr_list.append('%d' % time_last_seen[i,i])
            node_attr_list.append("1")
            node_attr_list.append("]")


            node_label_list.append("weight")
            node_label_list.append("count")
            node_label_list.append("intermittency")
            node_label_list.append("first_seen")
            node_label_list.append("last_seen")
            node_label_list.append("direct")
            node_label_list.append("")




        else:
            # nodeElem.appendChild(new_attribute('integer', 'direct', '0'))

            node_attr_list.append('0')
            node_attr_list.append("]")


            node_label_list.append("direct")
            node_label_list.append("")

    for i in range(TAG_ID_SPAN):
        for j in range(i+1,TAG_ID_SPAN):
            if not cum_adj[i, j]:
                continue

            edge_attr_list.append("[")
            edge_attr_list.append('%d' % i)
            edge_attr_list.append('%d' % j)
            edge_attr_list.append('"%d (pp) %d"' % (TAG_ID_MIN+i, TAG_ID_MIN+j) )



            edge_attr_list.append('%d' % cum_adj[i,j])
            edge_attr_list.append('%d' % onoff[i,j])
            edge_attr_list.append('%d' % (onoff[i,j] / float(cum_adj[i,j]) / 2.0) )
            edge_attr_list.append('%d' % time_first_seen[i,j])
            edge_attr_list.append('%d' % time_last_seen[i,j])
            edge_attr_list.append('%d' % (TAG_ID_MIN+i) )
            edge_attr_list.append('%d' % (TAG_ID_MIN+j) )

            edge_attr_list.append("")

            edge_label_list.append("edge")
            edge_label_list.append("source")
            edge_label_list.append("target")
            edge_label_list.append("label")



            edge_label_list.append("weight")
            edge_label_list.append("count")
            edge_label_list.append("intermittency")
            edge_label_list.append("first_seen")
            edge_label_list.append("last_seen")
            edge_label_list.append("node_A")
            edge_label_list.append("node_B")
            edge_label_list.append("]")


    # -----------------------------------



    nodes_all_together=zip(node_label_list,node_attr_list)

    n2=["%s %s" %(x,y) for x,y in nodes_all_together]

    n3="\n".join(n2)




    edges_all_together=zip(edge_label_list,edge_attr_list)


    e2=["%s %s" %(x,y) for x,y in edges_all_together]

    e3="\n".join(e2)

    print e3

    tot=n3+e3

    #if needed, change the value of "directed"

    ini="\n".join(["graph [", "directed 0",""])
    end="\n".join(["","]" ])

    tot=ini+tot+end

    return tot


#################################################################################################################



f = open("time_grid.dat")
time_grid = [map(int, string.split(line)) for line in f.readlines()]
f.close()

time_grid = s.array(time_grid, dtype="int64")


print ("time_grid is", time_grid )

loader = Loader(sys.argv[1:-1], experiment=EXPERIMENT, decode=DECODE, xxtea_crypto_key=XXTEA_CRYPTO_KEY, load_contacts=1, unique_contacts=1, load_sightings=0,
                 mapping=M,start_time=time_grid[0], stop_time=time_grid[-1])

# loader = Loader(sys.argv[1:], decode=0,experiment='OBG',load_contacts=1, unique_contacts=1, load_sightings=0, mapping=M)


# analyzer = ContactGraphAnalyzer(loader, DELTAT, get_graph=False, get_adj=True, tag_id_min=TAG_ID_MIN, tag_id_max=TAG_ID_MAX, isolated_nodes=True, weighted=False)

analyzer = ContactGraphAnalyzer(loader, DELTAT, get_graph=False, get_adj=True, tag_id_min=TAG_ID_MIN, tag_id_max=TAG_ID_MAX, isolated_nodes=True, weighted=True)



count_day=1

for frame in analyzer:

    frame_time = frame['time']

    if (frame_time > time_grid[count_day]):
        direct_indirect_seen = cum_adj.sum(0)

        total=collect_node_and_edge_properties(TAG_ID_SPAN,direct_indirect_seen,\
                                               TAG_ID_MIN,cum_adj,onoff,time_first_seen, time_last_seen)


        filename="graph_%01d"%(count_day)
        filename=filename+"_.gml"


        g=open(filename, "w")
        g.write(total)
        g.close()
        

        
        cum_adj = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
        onoff = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
        adj_prev = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
        time_first_seen = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
        time_last_seen = zeros((TAG_ID_SPAN, TAG_ID_SPAN), uint32)
        frame_count = 0

        count_day=count_day+1
 
    frame_count += 1
    adj = frame['adj']

    tags_high_k = set(where(adj.sum(0) > K_MAX)[0])
    tags_ignore = list(tags_high_k)
    adj[tags_ignore,:] = 0
    adj[:,tags_ignore] = 0

    cum_adj += adj
    onoff += where(adj_prev != adj, 1, 0)

    time_first_seen = where(onoff == 1, frame_time, time_first_seen)
    time_last_seen = where(adj > 0, frame_time, time_last_seen)

    adj_prev = adj

    print '#%d' % frame_count, time.ctime(frame['time'])





print "So far so good"