Commit ·
d5a85b5
1
Parent(s): a969444
Update in time format
Browse files- preprocessor.py +19 -9
preprocessor.py
CHANGED
|
@@ -1,11 +1,11 @@
|
|
| 1 |
import re
|
| 2 |
import pandas as pd
|
| 3 |
|
| 4 |
-
pattern_12 = '\d{1,2}\/\d{1,2}\/\d{2,4},\s\d{1,2}:\d{1,2}\s[a,m,p]+\s-\s'
|
| 5 |
-
pattern_24 = '\d{1,2}\/\d{1,2}\/\d{2,4},\s\d{1,2}:\d{1,2}\s-\s'
|
| 6 |
-
pattern_ph_num = '\d{1,2} \d{5} \d{5}'
|
| 7 |
|
| 8 |
def checker(data):
|
|
|
|
|
|
|
|
|
|
| 9 |
check_1 = re.search(pattern_12, data)
|
| 10 |
check_2 = re.search(pattern_24, data)
|
| 11 |
|
|
@@ -16,6 +16,10 @@ def checker(data):
|
|
| 16 |
|
| 17 |
|
| 18 |
def preprocess(data):
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
x = re.search(pattern_12, data)
|
| 20 |
|
| 21 |
# For 12 hour format
|
|
@@ -24,9 +28,12 @@ def preprocess(data):
|
|
| 24 |
dates = re.findall(pattern_12, data)
|
| 25 |
df = pd.DataFrame({'user_message': messages, 'message_date': dates})
|
| 26 |
|
| 27 |
-
|
| 28 |
-
df['message_date']
|
| 29 |
-
|
|
|
|
|
|
|
|
|
|
| 30 |
|
| 31 |
else:
|
| 32 |
messages = re.split(pattern_24, data)[1:]
|
|
@@ -34,10 +41,13 @@ def preprocess(data):
|
|
| 34 |
|
| 35 |
df = pd.DataFrame({'user_message': messages, 'message_date': dates})
|
| 36 |
|
| 37 |
-
|
| 38 |
-
df['message_date']
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
|
| 40 |
-
|
| 41 |
df.rename(columns={'message_date': 'date'}, inplace=True)
|
| 42 |
|
| 43 |
users = []
|
|
|
|
| 1 |
import re
|
| 2 |
import pandas as pd
|
| 3 |
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
def checker(data):
|
| 6 |
+
pattern_12 = '\d{1,2}\/\d{1,2}\/\d{2,4},\s\d{1,2}:\d{1,2}\s[a,m,p]+\s-\s'
|
| 7 |
+
pattern_24 = '\d{1,2}\/\d{1,2}\/\d{2,4},\s\d{1,2}:\d{1,2}\s-\s'
|
| 8 |
+
|
| 9 |
check_1 = re.search(pattern_12, data)
|
| 10 |
check_2 = re.search(pattern_24, data)
|
| 11 |
|
|
|
|
| 16 |
|
| 17 |
|
| 18 |
def preprocess(data):
|
| 19 |
+
pattern_12 = '\d{1,2}\/\d{1,2}\/\d{2,4},\s\d{1,2}:\d{1,2}\s[a,m,p]+\s-\s'
|
| 20 |
+
pattern_24 = '\d{1,2}\/\d{1,2}\/\d{2,4},\s\d{1,2}:\d{1,2}\s-\s'
|
| 21 |
+
pattern_ph_num = '\d{1,2} \d{5} \d{5}'
|
| 22 |
+
|
| 23 |
x = re.search(pattern_12, data)
|
| 24 |
|
| 25 |
# For 12 hour format
|
|
|
|
| 28 |
dates = re.findall(pattern_12, data)
|
| 29 |
df = pd.DataFrame({'user_message': messages, 'message_date': dates})
|
| 30 |
|
| 31 |
+
try:
|
| 32 |
+
df['message_date'] = pd.to_datetime(
|
| 33 |
+
df['message_date'], format='%d/%m/%y, %I:%M %p - ')
|
| 34 |
+
except ValueError:
|
| 35 |
+
df['message_date'] = pd.to_datetime(
|
| 36 |
+
df['message_date'], format='%d/%m/%Y, %I:%M %p - ')
|
| 37 |
|
| 38 |
else:
|
| 39 |
messages = re.split(pattern_24, data)[1:]
|
|
|
|
| 41 |
|
| 42 |
df = pd.DataFrame({'user_message': messages, 'message_date': dates})
|
| 43 |
|
| 44 |
+
try:
|
| 45 |
+
df['message_date'] = pd.to_datetime(
|
| 46 |
+
df['message_date'], format='%d/%m/%y, %H:%M - ')
|
| 47 |
+
except ValueError:
|
| 48 |
+
df['message_date'] = pd.to_datetime(
|
| 49 |
+
df['message_date'], format='%d/%m/%Y, %H:%M - ')
|
| 50 |
|
|
|
|
| 51 |
df.rename(columns={'message_date': 'date'}, inplace=True)
|
| 52 |
|
| 53 |
users = []
|