def compare_2ts(data, args):
col = ['blue', 'red', 'black', 'green']
lab = data[0].columns[-2:]
for i in range(len(data)):
data[i][lab] = standartization_ts(data[i][lab])
IO.plot_2ts(iocfg(x=data, label=lab, grid=True, figsize=(20,9), color=col))
print(data[0][lab[1]].corr(data[1][lab[1]]))
def ips_distribution(data, args):
default_ip_field_name = "clientip"
if args.verbose:
print(data)
df = pd.DataFrame(list(Counter(data[default_ip_field_name]).items()),
columns=['IP', 'Count']).sort_values(by='Count',
ascending=False,
na_position='last')
IO.write(df.dropna(axis='rows'), args, "_ips.csv")
IO.write(pd.DataFrame({default_ip_field_name: df['IP'].dropna()}), args,
"_ipflist.csv")
logging.debug("Df size is %d items.", len(df.index))
if args.verbose:
print(df)
df = IP.convert(df, args).dropna(axis='rows')
IO.plot(
iocfg(x=df['IP'],
y=df['Count'],
markersize=0.1,
color='b',
grid=True,
figsize=(20, 9),
title="IPs"))
IO.savefig(args, "_ips.eps")
def averages_check(_ip_avg, args):
logging.debug("Plot average_freq_interval")
# plot average spectre:
_ip_avg = _ip_avg.drop(['noname'], axis=1)
means, sd = average_freq_interval(_ip_avg, args)
spd = find_spectre_deviation(_ip_avg, means, sd, args)
for i in range(int(args.state['time']['avgTop'])):
if args.state['time']['avgType'] == 'reverse':
idx = len(spd.index) - i - 1
else:
idx = i
args.output = "_".join(["ms", spd['ip'][idx], args.state['time']['avgType'],
args.tb[0], args.tb[1]])
par = iocfg(
x = means.index, y=means, addition_y=sd,
figsize = (20, 9), grid=True,
title="Mean spectre from {} to {} freq (by {} points, {} sample)".format(high_b, low_b,
len(_ip_avg.columns), args.round),
compared = _ip_avg[spd['ip'][idx]]
)
IO.plot_line_err(par)
IO.savefig(args, ".eps")
args.output = "_".join(["ip", args.tb[0], args.tb[1]])
IO.Time.write_pretty_txt(pd.DataFrame.from_dict({"IP": ip, "Freq": freqs}), args)
par = iocfg(x = iplist[iplist.columns[1]][:i],
y = iplist[iplist.columns[2]][:i],
title = "IP freq distribution",
figsize = (12, 12),
addition_x = idx,
addition_y = iplist[iplist.columns[2]][idx]
)
IO.plot_line(par)
IO.savefig(args, "_ip_distr.eps")
def plot_entropy_distance(self, ds, dist):
self._args.output = "{}_".format(ds.index[0].time())
etype = self._args.state['math']['entropyType']
if self._args.state['io']['epsOut'] == 'yes':
col = ['blue', 'red']
IO.plot_general_ts(
iocfg(x=[ds, dist],
grid=True,
title="{} entropy for {}".format(etype,
self._args.field),
figsize=(20, 9),
color=col), self._args)
IO.savefig(self._args, "{}_entropy.eps".format(etype))
IO.write(ds, self._args, "{}_entropy.csv".format(etype))
IO.write(dist, self._args, "{}_entropy_dist.csv".format(etype))
def groupby_ops(data, args):
c = Counter(data[args.category])
data = data.groupby(args.category)[args.field].median().to_frame()
data = data.reset_index()
res = len(data.index) * [None]
res_c = len(data.index) * [None]
col = len(data.index) * [None]
for i in range(len(data.index)):
res[i] = int(ipaddress.IPv4Address(data[args.category][i]))
res_c[i] = c[data[args.category][i]]
col = color_assign(res_c[i])
data = data.join(pd.DataFrame({'IP': res, 'Count': res_c}))
IO.plot(
iocfg(x=data['IP'],
y=data[args.field],
markersize=1.5,
color=col,
grid=True,
figsize=(13, 13)))
IO.write(data.sort_values(by=[args.field], ascending=False), args,
"_groupby.csv")