def disconnect(api_id=api_id,
api_hash=api_hash,
client=None,
session_name=default_session_name) -> NoReturn:
"""Ensure previous client session is closed."""
print("Trying to end the previous client session...")
try:
if client is None:
client = TelegramClient(session_name, api_id, api_hash)
client.log_out()
print('Ended previous client session')
except ConnectionError:
print('Connection error raised: already disconnected from Telegram')
except:
raise
print("Disconnected from Telegram")
class TelegramDialogSource(DialogSource):
def __init__(self, api_id: int, api_hash: str):
self._client = TelegramClient(session=None,
api_id=api_id,
api_hash=api_hash).start()
def __enter__(self):
return self
def __exit__(self, ecx_type, ecx_value, traceback):
self._client.log_out()
self._client.disconnect()
def getDialogs(self) -> Iterable[Dialog_]:
for dialog in self._client.iter_dialogs():
yield TelegramDialog(dialog)
print("\n\nCOUNTED COMPLETED! OPTIONS:")
while True:
print(
"\nDo you want to log out of TLCounter? If you want to count your messages frequently, you might want to keep your session logged in."
)
print("> Available commands: ")
print(" !1: Log out")
print(" !Q: Close the program without logging out.")
print()
answer = str(input("Enter a command: "))
answer = answer.replace(" ", "")
answer = answer.upper()
if not (answer == '!Q' or answer == '!1'):
while True:
print()
answer = input(
"The command you entered was not valid. Please, enter a valid one: "
)
answer.replace(" ", "")
answer.upper()
if (answer == "!Q") or (answer == "!1"):
break
if (answer == "!Q"):
client.disconnect()
exit(0)
if (answer == "!1"):
print("Logging you out of Telegram...")
client.log_out()
getpass("Done! Press ENTER to close TLCounter!: ")
exit(0)
file = client1.send_file(ChosenChat, "DB.aes")
print("\n\nWaiting for a response...")
client1.run_until_disconnected()
print(
"Your partner has been authorised in Telegram successfully!\nYou can keep this window open for revoking the session in case you think it's necessary.\nYou can close this app if you don't want to revoke it.\n\n"
)
getpass("\n\nPress ENTER to revoke the session: ")
print("\nLogging out...")
while not client1.is_connected():
client1.connect()
if client1.is_connected():
break
else:
print("There was a connection problem... Retrying in 2s")
time.sleep(2)
client1.log_out()
print("\n\nLogged out!")
else:
getpass("There is already a session logged in. Press ENTER to revoke it: ")
while not client1.is_connected():
client1.connect()
if client1.is_connected():
break
else:
print("There was a connection problem... Retrying in 2s")
time.sleep(2)
client1.log_out()
try:
os.remove("DB.aes")
except:
pass
class TelegramAnalytics:
"""
A that extracts information about messages times.
"""
def __init__(self, api_id, api_hash):
self.__api_id = api_id
self.__api_hash = api_hash
self.__from_zone = tz.tzutc()
self.__to_zone = tz.tzlocal()
self.__client = TelegramClient('session_name', api_id, api_hash)
self.__client.start()
def __del__(self):
self.__client.log_out()
self.__client.disconnect()
def get_all_message_date_times(self, username, limit=10):
"""
Get dates and times for all messages exchanged between you and an entity.
:param str username: Username (or phone number) of entity to get chat history from.
:param int limit: The number of messages to extract information from.
:return: Array of dates and times that each messages was sent.
:rtype: [datetime.datetime]
>>> ca = TelegramAnalytics( api_id, api_hash )
>>> ca.get_all_message_date_times('+11234567891',limit=1)
>>> ca.get_all_message_date_times('quartz_husky',limit=1)
"""
return list(
map(lambda x: x.date.astimezone(self.__to_zone),
self.__client.iter_messages(username, limit=limit)))
def get_all_message_times(self, username, limit=10):
"""
Get time stamps (hours, minutes, seconds, microseconds)
for all messages exchanged between you and an entity.
:param str username: Username (or phone number) of entity to get chat history from.
:param int limit: The number of messages to extract information from.
:return: Array of times that each messages was sent.
:rtype: [datetime.time]
>>> ca = TelegramAnalytics( api_id, api_hash )
>>> ca.get_all_message_date_times('+11234567891',limit=1)
>>> ca.get_all_message_date_times('quartz_husky',limit=1)
"""
return list(
map(lambda x: x.date.astimezone(self.__to_zone).time(),
self.__client.iter_messages(username, limit=limit)))
def message_time_histogram(self, username, limit=10):
"""
Get a time histrogram for all messages exchanged between you and an entity.
:param str username: Username (or phone number) of entity to get chat history from.
:param int limit: The number of messages to extract information from.
>>> from fnw_client import TelegramAnalytics
>>> import matplotlib.pyplot as plt
>>>
>>> plt.figure()
>>>
>>> ca = TelegramAnalytics( api_id, api_hash )
>>> time_historgram = ca.message_time_histogram('quartz_husky',limit=100)
>>> time_histogram.plot(kind='bar')
>>>
>>> plt.show()
"""
times = get_all_message_date_times(username, limit)
df = pd.DataFrame({'hour': times})
return df.groupby(df['hour'].dt.hour).count()
def __text_sentiment(self, text):
def clean_text(self, text):
return ' '.join(
re.sub("(@[A-Za-z0-9]+)|([^0-9A-Za-z \t])|(\w+:\/\/\S+)", " ",
text).split())
cln_text = clean_text(text)
return None if not cln_text else TextBlob(cln_text).sentiment.polarity
def get_sentiments(self, username, limit=10):
message_iterator = self.__client.iter_messages(username, limit=limit)
return [
self.__text_sentiment(x.text) for x in message_iterator
if isinstance(x.text, str)
]
def message_sentiment_histogram(self, username, limit=10):
"""
Get a sentiment histrogram for all messages exchanged between you and an entity.
"""
sentiments = get_sentiments(username, limit)
df = pd.DataFrame({'sentiment': sentiments})
# Remove zero values and NAs
df_nz = df[(df.T != 0).any()].dropna()
return df_nz.hist(bins=20)
def message_sentiment_rolling_avg(self, username, limit=10):
"""
Get rolling average data of all messages exchanged between you and an entity.
"""
sentiments = get_sentiments(username, limit)
df = pd.DataFrame({'sentiment': sentiments})
# Remove zero values and NAs
df_nz = df[(df.T != 0)].any().dropna()
rolling_avg_factor = int(df_nz.shape[0] / 100) + 1
return df_nz.rolling(rolling_avg_factor).mean()
def message_sentiment_vs_time(self, username, limit=10):
"""
Get a graph of sentiment as a function of hour of the day.
:param str username: Username (or phone number) of entity to get chat history from.
:param int limit: The number of messages to extract information from.
:return: A tuple with the following format: (hour,sentiment,variance-in-sentiment)
:rtype: (int,float,float)
>>> from fnw_client import TelegramAnalytics
>>> import matplotlib.pyplot as plt
>>>
>>> plt.figure()
>>>
>>> ca = TelegramAnalytics( api_id, api_hash )
>>> sentiment_and_time = ca.message_sentiment_and_time('quartz_husky',limit=100)
>>> variances = [ x.variance() for x in sentiment_and_time ]
>>>
>>> plt.bar(range(len(sentiment_and_time)), [val.sentiment() for val in sentiment_and_time], align='center',yerr=variances)
>>> plt.xticks(range(len(sentiment_and_time)), [val.time() for val in sentiment_and_time])
>>> plt.xticks(rotation=70)
>>> plt.xlabel('Hour')
>>> plt.ylabel('Average Sentiment')
>>>
>>> plt.show()
"""
class TimeAndSentiment:
def __init__(self, time, sentiment):
self.__time = time
self.__sentiment = sentiment
def time(self):
return self.__time
def sentiment(self):
return self.__sentiment
def get_sentiments_and_times(self, username, limit=10):
message_iterator = self.__client.iter_messages(username,
limit=limit)
get_datetime = lambda x: x.date.astimezone(self.__to_zone).time()
get_sentiment = lambda x: self.__text_sentiment(x.text)
ts = [
TimeAndSentiment(get_datetime(x), get_sentiment(x))
for x in message_iterator if isinstance(x.text, str)
]
# Sorted output by hours, which is necessary for groupby to work properly
return sorted(ts, key=lambda x: x.time().hour)
hour_group = groupby(get_sentiments_and_times(username, limit),
lambda x: x.time().hour)
average_sentiment = lambda group: mean(
[x.sentiment() for x in group if x.sentiment() is not None])
def stdev_sentiment(group):
clean_group = [x for x in group if x.sentiment() is not None]
return stdev([x.sentiment() for x in clean_group
]) if (len(clean_group) >= 2) else 0
return [
TimeSentimentVariance(key, average_sentiment(group),
stdev_sentiment(group))
for key, group in hour_group
]
