forked from stephentu/gchatviz
/
analyze.py
271 lines (222 loc) · 7.41 KB
/
analyze.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
#!/usr/bin/env python
from message import message
import argparse
import pickle
import sys
import pytz
import os
import errno
import itertools as it
import numpy
import csv
import ntpath
import hashlib, uuid
from datetime import datetime as dt
from datetime import time as dtime
from bs4 import BeautifulSoup
import matplotlib.pyplot as plt
from matplotlib.dates import YearLocator, MonthLocator, \
HourLocator, MinuteLocator, DateFormatter
def mkdirp(path):
# worst idiom ever
try:
os.makedirs(path)
except OSError as exc:
if exc.errno == errno.EEXIST:
pass
else:
raise
def argmax(seq, keyfn=lambda x: x):
best, best_value = seq[0], keyfn(seq[0])
for e in seq[1:]:
if keyfn(e) > best_value:
best, best_value = e, keyfn(e)
return best
def groupby(dates):
"""
given a list of ordered objects, returns a list of [(obj, count)],
in ascending object order
"""
m = {}
for d in dates:
m[d] = m.get(d, 0) + 1
return sorted(m.iteritems())
def write_messages(messages, outfile, anonymize, fromsender):
uids = {}
sorted_bydate = sorted(messages, key=lambda m: m._date)
with open(outfile, "w") as out:
writer = csv.writer(out)
writer.writerow(["sender","message","datetime"])
for m in sorted_bydate:
if anonymize:
mess = get_hidden(m._message) # hide message
if fromsender in m._fromuser: # hide sender (unless chat is authored by fromsender)
author = "me"
else:
author = get_hash(m._fromuser, uids)
else:
mess = m._message
author = m._fromuser
row = [author, mess.replace('\n', ' '), clean_date(str(m._date))]
writer.writerow(row)
# Returns a unique ID for the user
# uids: a dictionary of existing users to ids
def get_hash(user, uids):
if user in uids.keys():
return uids[user]
elif not uids.keys():
uids[user] = 1
return 1
else:
new_uid = numpy.max([int(key) for key in uids.values()]) + 1
uids[user] = new_uid
return new_uid
# Returns a message in anonymized form: i.e. "xxx xxxxxxxxx xx"
def get_hidden(message):
import re
hidden = re.sub('\S','x', message)
return hidden
def write_stats(stats, outfile):
with open(outfile, "w") as out:
out.write(str(stats))
def clean_date(date):
without_microseconds = date.split('.')[0]
time = without_microseconds.split(" ")[1]
if "-" in time:
to_remove = time.split("-")[1]
return without_microseconds.replace("-" + to_remove, '')
else:
return without_microseconds
def getstats(messages, sender = ""):
# chats per day
dateinfo = [m._date.date() for m in messages]
ts = groupby(dateinfo)
msgcounts = [x[1] for x in ts]
stats = {}
stats["chats"] = {"avg": numpy.average(msgcounts), "min": numpy.min(msgcounts), "max": numpy.max(msgcounts)}
# chats per day from a particular sender
if sender:
fromsender = filter_bysender(messages, sender)
dateinfo_fromsender = [m._date.date() for m in fromsender]
ts_fromsender = groupby(dateinfo_fromsender)
msgcounts_fromsender = [x[1] for x in ts_fromsender]
stats["chats_fromme"] = {"avg": numpy.average(msgcounts_fromsender), "min": numpy.min(msgcounts_fromsender), "max": numpy.max(msgcounts_fromsender)}
return stats
def filter_bysender(messages, sender):
filtered = []
for m in messages:
if sender in m._fromuser:
filtered.append(m)
return filtered
def plot(messages, outfolder):
# count messages sent by user
msg_counts = {}
word_counts = {}
for m in messages:
nwords = len(m._message.split())
msg_counts[m._fromuser] = msg_counts.get(m._fromuser, 0) + 1
word_counts[m._fromuser] = word_counts.get(m._fromuser, 0) + nwords
dateinfo = [m._date.date() for m in messages]
fig = plotbydate(dateinfo)
fig.savefig(os.path.join(outfolder, 'activity-date.pdf'))
fig = plotbyavgtime([m._date for m in messages])
fig.savefig(os.path.join(outfolder, 'activity-time.pdf'))
def plotbydate(dates):
ts = groupby(dates)
#print argmax(ts, lambda x: x[1])
years = YearLocator()
months = MonthLocator()
yearsFmt = DateFormatter('%Y')
fig, ax = plt.subplots()
ax.plot_date([t[0] for t in ts], [t[1] for t in ts], '-')
# format the ticks
ax.xaxis.set_major_locator(years)
ax.xaxis.set_major_formatter(yearsFmt)
ax.xaxis.set_minor_locator(months)
ax.autoscale_view()
# format the coords message box
ax.fmt_xdata = DateFormatter('%Y-%m-%d')
ax.grid(True)
fig.autofmt_xdate()
ax.set_ylabel('Messages/Day')
ax.set_title('GChat Activity')
return fig
def buckettime(t):
return t.replace(minute=int(t.minute / 10.0)*10, second=0, microsecond=0)
def plotbyavgtime(datetimes):
ndays = len(set(d.date() for d in datetimes))
# bucket each time object
datetimes = map(buckettime, datetimes)
counts = {}
for k, g in it.groupby(datetimes):
mykey = k.time()
mysum, mycnt = counts.get(mykey, (0,0))
counts[mykey] = (mysum + len(list(g)), mycnt + 1)
ts = []
# plot can only handle date objects, not time objects
hack = dt.now()
for h in xrange(24):
for m in xrange(0, 60, 10):
t = dtime(h, m)
mysum, mycnt = counts.get(t, (0,0))
assert mycnt <= ndays
ts.append((dt.combine(hack, t), mysum / float(ndays)))
print argmax(ts, lambda x: x[1])
timefmt = DateFormatter('%I:%M%p')
fig, ax = plt.subplots()
ax.plot([t[0] for t in ts], [t[1] for t in ts], '-')
# format the ticks
ax.xaxis.set_major_locator(HourLocator())
ax.xaxis.set_major_formatter(timefmt)
ax.xaxis.set_minor_locator(MinuteLocator(interval=10))
ax.autoscale_view()
# format the coords message box
ax.fmt_xdata = DateFormatter('%I:%M%p')
ax.grid(True)
fig.autofmt_xdate()
ax.set_ylabel('avg msgs/day at time')
ax.set_title('GChat Activity Breakdown by Time of Day')
return fig
def path_leaf(path):
head, tail = ntpath.split(path)
return tail or ntpath.basename(head)
if __name__ == '__main__':
p = argparse.ArgumentParser()
p.add_argument('--infolder', required=True)
p.add_argument('--outfolder', required=True)
p.add_argument('--msg', action='store_true')
p.add_argument('--anonymize', action='store_true') # additional option for --msg
p.add_argument('--plot', action='store_true')
p.add_argument('--stats', action='store_true')
p.add_argument('--fromsender', required=True)
args = p.parse_args()
mkdirp(args.outfolder)
print "grabbing messages..."
d = []
for infile in os.listdir(args.infolder):
with open(os.path.join(args.infolder, infile), 'r') as infp:
d += pickle.load(infp)
# XXX: hacky-- add timezone info
tz = pytz.timezone('US/Pacific')
for idx in xrange(len(d)):
if d[idx]._date.tzinfo is None:
d[idx]._date = d[idx]._date.replace(tzinfo=tz)
# sort by time
messages = sorted(d, key=lambda x: x._date)
if args.msg:
print "processing message data..."
if args.anonymize:
write_messages(d, os.path.join(args.outfolder, "messagedata-anon.csv"), True, args.fromsender)
write_messages(d, os.path.join(args.outfolder, "messagedata.csv"), False, args.fromsender)
if args.plot:
print "plotting..."
plot(messages, args.outfolder)
if args.stats:
print "getting statistics..."
stats = getstats(messages, args.fromsender)
print stats
write_stats(stats, os.path.join(args.outfolder, "stats"))
# write to text file
#with open(outfile, 'w') as outfp:
# for m in d:
# print >>outfp, '%s -> %s (%s): %s' % (m._fromuser, m._touser, m._date.strftime('%x %X'), m._message)