def test_strip11s_too_shorts(self):
r = HepReader()
self.assertEqual(r._strip11s("0011999", date(2000, 11, 1)), "11999")
self.assertEqual(r._strip11s("11999", date(2000, 11, 1)), "11999")
self.assertEqual(r._strip11s("1199", date(2000, 1, 1)), "1199")
self.assertEqual(r._strip11s("0112999", date(2001, 12, 1)), "112999")
def test_strip11s_cross_cits(self):
r = HepReader()
self.assertEqual(r._strip11s("119807999", date(1998, 07, 1)), "9807999")
self.assertEqual(r._strip11s("11107999", date(2001, 07, 1)), "107999")
self.assertEqual(r._strip11s("11207999", date(2002, 07, 1)), "207999")
self.assertEqual(r._strip11s("1112999", date(2000, 12, 1)), "12999")
self.assertEqual(r._strip11s("117999", date(2000, 07, 1)), "7999")
def plot_community_size_distribution():
"""Plot overall distribution for HEP-formatted files"""
sizes = Counter()
all_sizes = []
for year in range(1992, 2004):
hp = HepReader.get_for_year(year)
rm = RoleMining(hp.get_nodes(), hp.get_edges())
sizes_of_communities = [len(c) for c in rm.communities.values()]
all_sizes.extend(sizes_of_communities)
counted = Counter(sizes_of_communities)
sizes += counted
min_size = 10
max_size = max(all_sizes)
year = "overall"
P.xlabel("Size of community [members]")
P.ylabel("Number of communities")
P.suptitle("Year {}, communities bigger than 10 members\nNon-overlapping community size distribution".format(year))
n, bins, patches = P.hist(filter(lambda x: x > min_size, all_sizes))
autolabel(patches)
P.xticks([min_size] + range(min_size-10, max_size + 1, 100) )
P.xlim(min_size, max_size)
P.show()
def plot_community_size_distribution_from_cfinder(year, k, prop, color):
"""Plot community size distribution from CFinder cliques output files"""
filename = "datasets/enron/communities/8-{0}/k={1}/{2}".format(year, k, prop)
# filename = "datasets/hepth/communities/cit-HepTh-{0}/k={1}/{2}".format(year, k, prop)
lines = HepReader.read_lines(filename)
sizes = {int(size): int(count) for size, count in [line.split() for line in lines if len(line) != 0]}
x = sizes.keys()
y = sizes.values()
P.bar(x, y, label="year:{}, k={}, max={}".format(year, k, max(x)), align='center', alpha=0.7, color=color)
P.xlabel("Size of community [members]")
P.ylabel("Number of communities")
def test_date_from_id_regular(self):
self.assertEqual(HepReader._date_from_id("212999"), date(2002, 12, 01))
def test_date_from_idXXcent(self):
self.assertEqual(HepReader._date_from_id("9901999"), date(1999, 01, 01))
def test_date_from_id_no_zeros(self):
self.assertEqual(HepReader._date_from_id("1999"), date(2000, 01, 01))
def _get_edges_per_slot():
slots = {} # {year : edges}
for year in range(1992, 2004):
slots[year] = HepReader.read_edges("datasets/hepth/timeslots/cit-HepTh-{0}.edges".format(year))
return slots
P.xlim(min_size, max_size)
P.show()
def plot_data_distribution():
"""Plot number of new papers and new citations per year, take Hep files as input"""
minyear = 1992
maxyear = 2003
x = [datetime(year, 01, 01) for year in range(minyear, maxyear + 1)]
x.extend([datetime(year, 06, 01) for year in range(minyear, maxyear + 1)])
x.extend([datetime(year, 9, 1) for year in range(minyear, maxyear + 1)])
x.extend([datetime(year, 12, 1) for year in range(minyear, maxyear + 1)])
x = sorted(x)
dates = HepReader.read_dates("/home/stpk/dev/role-mining/datasets/hepth/cit-HepTh-dates-cleaned.txt")
edges = HepReader.read_edges("/home/stpk/dev/role-mining/datasets/hepth/cit-HepTh.txt")
slots = HepReader.split_to_timeslots(dates, edges, x)
for slot in slots.values():
print len(slot)
citats = [len(slots[t]) for t in x]
print citats
print x
fig, ax = P.subplots()
ax.xaxis_date()
P.title("Nowe publikacje na kwartal")
P.plot(x, citats, color='b', alpha=0.5, label="Nowe cytowania")
ax.set_xticks(x)
P.xticks(rotation=70)
