def __init__(self):
self.scales = [(_("Hour"),hours(1),MinuteLocator(interval=10)),
(_("Half-hour"),minutes(30),MinuteLocator(interval=5)),
(_("10-Minute"),minutes(10),MinuteLocator(interval=2)),
(_("5-Minute"),minutes(5),MinuteLocator(interval=1)),
(_("2-Minute"),minutes(2),SecondLocator(interval=20)),
(_("Minute"),minutes(1),SecondLocator(interval=10)),
(_("30-Seconds"),minutes(0.5),SecondLocator(interval=5)),
(_("12-Seconds"),minutes(0.2),SecondLocator(interval=2))
]
self.cur = 0
self.pos = None
def time_truncate(dictionary, start_time, end_time):
""" take a dictionary and reduce it to the desired interval.
We do this rather than integrate it into time_to_tweets to keep
time_to_tweets from becoming more complicated; this doesn't add
any extra time, and requires only slightly more memory. It could be
abstracted even more and called in time_to_tweets, but this lets us keep
the logic separate and easier to debug.
"""
# We need to preprocess times into our desired mdates format.
zero_time = list(dictionary.keys())[0]
start_time = start_time.split(':')
start_time = mdate.hours(float(start_time[0])) + \
mdate.minutes(float(start_time[1]))
start_time = start_time + zero_time
end_time = end_time.split(':')
end_time = mdate.hours(float(end_time[0])) + \
mdate.minutes(float(end_time[1]))
end_time = zero_time + end_time
# we need a new dictionary. It costs a bit of memory.
new_dict = OrderedDict({})
for time in dictionary:
if time < start_time:
continue
elif time > start_time and time < end_time:
new_dict[time] = dictionary[time]
elif time > end_time:
break
else:
break
return new_dict
def time_to_tweets(filename, search_list, interval=15):
"""
This determines how to bucket the times for a line plot, and returns a
dictionary.
"""
fi = open(filename, 'r', encoding='utf_8')
time_dictionary = OrderedDict({})
interval_time = mdate.minutes(interval)
# this is set to zero because of the way initialization works later
curr_time = 0
for line in nonblank_lines(fi):
jdata = json.loads(line)
for term in search_list:
if term in jdata['text'].lower():
timestamp = datetime.datetime.strptime(jdata['created_at'], '%a %b %d \
%H:%M:%S %z %Y')
timename = mdate.date2num(timestamp)
# initializes the dictionary. This is true only once, but we do
# it here, rather than earlier, since it lets us initialize it
# for the first timename (which we won't know until we've
# started searching the JSON file).
if time_dictionary == {}:
time_dictionary[timename] = dict(search_list)
time_dictionary[timename][term] += 1
curr_time = timename
# this conditional should NOT be an elif!
if timename < curr_time + interval_time:
time_dictionary[curr_time][term] += 1
elif timename > curr_time + interval_time:
curr_time = timename
time_dictionary[curr_time] = dict(search_list)
time_dictionary[curr_time][term] += 1
fi.close()
return time_dictionary
