# and M[i] returns all measurements on metric i
pca = g.PCA(M)
pca.plot()
### Make network PCA with partitions ###
dl = g.DownloadGmaneData('~/.gmane2/')
dl.downloadedStats() # might take a while
print("made liststats")
lm = g.LoadMessages(dl.lists[0][0], basedir="~/.gmane2/")
print("loaded messages")
ds = g.ListDataStructures(lm)
print("made datastructures")
iN = g.InteractionNetwork(ds)
print("made interaction network")
nm = g.NetworkMeasures(iN)
print("network mesaures")
# Make PCA of degree, betweenness and clustering
# with in, out and total degree and strengths
# and with symmetry related measures
npca = g.NetworkPCA(nm)
# Make network partitioning
np = g.NetworkPartitioning(nm, 3, "g")
print("partitioned network")
# Plot PCA with partitions
npca = g.NetworkPCA(nm, np)
PDIR="pickledir/"
NEs=[] # for evolutions of the networks
for lid in dl.downloaded_lists[:1]:
label=labels[lid]
NEs.append(pRead("{}neP{}.pickle".format(PDIR,label)))
print(label+"{0:.2f} for PICKLE loading evolved PCA structures".format(T.time()-TT)); TT=T.time()
# NEs[0].network_measures[X] have E, N, degrees, etc and my be ok to make network partition
ne=NEs[0] # snapshots ordenados
nm=ne.networks_measures[-1]
# nao achei network partitionings no ne (binario de versao antiga da classe?)
# fazendo o partitionings
np=g.NetworkPartitioning(nm)
p.figure(figsize=(10.,3.))
p.subplots_adjust(left=0.08,bottom=0.18,right=0.99,top=0.87)
p.bar([i[0]-0.5 for i in np.bins],
np.empirical_distribution,
alpha=.7, label="empirical distribution")
#p.bar(n.log([i[0]-0.5 for i in np.bins]),
# n.log( np.empirical_distribution),
# alpha=.7, label="empirical distribution")
centers=[i[0] for i in np.bins]
binomial_probs=[]
for i, bin_ in enumerate(np.bins):
llimit=bin_[0]
rlimit=bin_[1]
dreload(g, exclude="pytz")
dl = g.DownloadGmaneData('~/.gmane2/')
dl.downloadedStats() # might take a while
print("made liststats")
lm = g.LoadMessages(dl.lists[0][0], basedir="~/.gmane2/")
print("loaded messages")
ds = g.ListDataStructures(lm)
print("made datastructures")
iN = g.InteractionNetwork(ds)
print("made interaction network")
nm = g.NetworkMeasures(iN)
print("network mesaures")
np = g.NetworkPartitioning(nm)
print("partitioned network")
np2 = g.NetworkPartitioning(nm, 2)
print("partitioned network")
np3 = g.NetworkPartitioning(nm, 3)
print("partitioned network")
np_ = g.NetworkPartitioning(nm, 1, "g")
print("partitioned network")
np2_ = g.NetworkPartitioning(nm, 2, "g")
print("partitioned network")
np3_ = g.NetworkPartitioning(nm, 3, "g")
print("partitioned network")
ps = [np, np2, np3] # partition by strength
pg = [np_, np2_, np3_] # partition by degree
print("{0:.2f} for statistics along time".format(T.time() - TT))
TT = T.time()
tss.append(ts)
pDump(ts, "{}ts{}.pickle".format(PDIR, lid))
iN = g.InteractionNetwork(ds)
print("made interaction network")
iNs.append(iN)
pDump(iN, "{}iN{}.pickle".format(PDIR, lid))
nm = g.NetworkMeasures(iN, exclude=["rich_club"])
print("network mesaures")
nms.append(nm)
pDump(nm, "{}nm{}.pickle".format(PDIR, lid))
np2_ = g.NetworkPartitioning(nm, 2, "g")
print("partitioned network")
nps.append(np2_)
pDump(nm, "{}np{}.pickle".format(PDIR, lid))
data_ = []
count = 0
for lid in dl.lists[4:6]:
lid = lid[0]
ds = dss[count]
np = nps[count]
count += 1
date1 = ds.messages[ds.message_ids[0]][2].isoformat().split("T")[0]
date2 = ds.messages[ds.message_ids[-1]][2].isoformat().split("T")[0]
N = ds.n_authors
Ns = [len(i) for i in np.sectorialized_agents__]