elif sys.argv[1] == "b":
corpus_type = "binary"
else:
corpus_type = "bow"
if len(sys.argv) <= 2:
topics_count = 3
else:
topics_count = int(sys.argv[2])
if len(sys.argv) <= 3:
src = "pp_reuters"
else:
src = sys.argv[3]
dtw = name.get_output_dir(corpus_type, topics_count, src)
t_1 = dtw + "/topics"
t_list1 = topics_io.read_topics(t_1)
bha_list = stl.bha_distance(t_list1, t_list1)
cos_list = stl.cos_distance(t_list1, t_list1)
kl_list = stl.kl_divergence(t_list1, t_list1)
jlist = stl.jaccard(t_list1, t_list1, 500)
kprepare = []
for topic in t_list1:
topic.sort()
kprepare.append(topic.list_words())
from scipy import stats
import sys
import utils.name_convention as namecon
from similarity.SimTopicLists import SimTopicLists
tclist, tctlist, wnlist = [],[],[]
wn_names = ["path", "wup", "lch", "res", "lin", "jcn"]
for n in range(len(wn_names)):
wnlist.append([])
for src in ["pp_reuters", "pp_brown"]:
for corpus_type in ["tfidf", "bow", "binary"]:
for topics_count in [5,10,15, 20]:
dname = namecon.get_output_dir(corpus_type, topics_count, src)
subtclist = []
ofile = open(dname + "/top_topics_20_start0.txt")
for line in ofile:
if "topic" in line:
subtclist.append(("tc"+src+corpus_type+str(topics_count)+line.split()[1],float(line.split()[2]), int(line.split()[1])))
subtclist = list(sorted(subtclist, key=lambda x: x[2]))
tclist.extend(subtclist)
subtctlist = []
ofile = open(dname + "/top_topics_tfidf_20.txt")
for line in ofile:
if "topic" in line:
subtctlist.append(("tct"+src+corpus_type+str(topics_count)+line.split()[1],float(line.split()[2]), int(line.split()[1])))
subtctlist = list(sorted(subtctlist, key=lambda x: x[2]))
tctlist.extend(subtctlist)