This package detects specific types of anomalies with an emphasis in looking for cumulative changes.
This package can be installed through PyPi using
pip install canomaly
or
pip3 install canomaly
>>> import pandas as pd
>>> from canomaly.searchtools import cumrexpy
>>> # Get some data
>>> data = {
'date': [
'2018-11-20',
'2018-11-21',
'2018-11-22',
'2018-11-22',
'2018-11-23',
'2018-11-24'],
'email': [
'john.doe@example.com',
'jane.smith@example.com',
'bob-johnson_123@example.com',
'sarah@mydomain.co.uk',
'frank@mydomain.com',
'jessica_lee@mydomain.com'
]
}
>>> df = pd.DataFrame(data)
>>> df['date'] = pd.to_datetime(df['date'])
>>> # Take a peek at the data
>>> df
date email
0 2018-11-20 john.doe@example.com
1 2018-11-21 jane.smith@example.com
2 2018-11-22 bob-johnson_123@example.com
3 2018-11-22 sarah@mydomain.co.uk
4 2018-11-23 frank@mydomain.com
5 2018-11-24 jessica_lee@mydomain.com
>>> # Extract regular expressions
>>> cumrexpy(df, 'email', 'date')
date
2018-11-20 [^john\.doe@example\.com$]
2018-11-21 [^[a-z]{4}\.[a-z]{3,5}@example\.com$]
2018-11-22 [^[a-z]{4,5}[.@][a-z]+[.@][a-z]+\.[a-z]{2,3}$,...
2018-11-23 [^frank@mydomain\.com$, ^[a-z]{4,5}[.@][a-z]+[...
2018-11-24 [^frank@mydomain\.com$, ^[a-z]+[.@_][a-z]+[.@]...
Name: email_grouped, dtype: objectWe can look at the results in markdown for clarity.
| date | email_grouped |
|---|---|
| 2018-11-20 00:00:00 | ['^john\.doe@example\.com$'] |
| 2018-11-21 00:00:00 | ['^[a-z]{4}\.[a-z]{3,5}@example\.com$'] |
| 2018-11-22 00:00:00 | ['^[a-z]{4,5}[.@][a-z]+[.@][a-z]+\.[a-z]{2,3}$', '^bob\-johnson_123@example\.com$'] |
| 2018-11-23 00:00:00 | ['^frank@mydomain\.com$', '^[a-z]{4,5}[.@][a-z]+[.@][a-z]+\.[a-z]{2,3}$', '^bob\-johnson_123@example\.com$'] |
| 2018-11-24 00:00:00 | ['^frank@mydomain\.com$', '^[a-z]+[.@_][a-z]+[.@][a-z]+\.[a-z]{2,3}$', '^bob\-johnson_123@example\.com$'] |
The documentation is available https://canomaly.readthedocs.io/en/latest/index.html, or you can build it locally using the following:
cd /path/to/canomaly/docs
make html