wombatoo 1.0.0

A Data-Ops tool for devs

WOMBATOO

Wombatoo is a python library which have the following features:

1. Dataframe summary: Summary of dataframe in CSV format
2. Datacard: A single document having Timestamp, data owner, data source, data refresh cycle, data details, Intended use, data summary, ethical considerations, Recommendation and caveats. Some fields are pre-populated and the user can populate the rest to define a consice contract between data owner and data user.
3. Version: a local versioning tool using which a user can use to version small to medium datasets locally.
4. Report: A pandas-profiling report which can be saved in HTML, JSON and PDF formats.
5. Schema: A csv document having the schema of the dataset. Users can use it to validate the schema in the data pipeline.

Installation

pip install wombatoo

from wombatoo import wombatoo
import pandas as pd

#reading df
df = pd.read_csv('base.csv')

if __name__ =="__main__":
    summary = wombatoo.metric_summary(df,path_to_save='summary.csv')
    datacard = wombatoo.datacard(df, path_to_save='Datacard_user.txt')
    version = wombatoo.version(df.head(5),file_name='fraud_bool', folder_path='./local_repo')
    report = wombatoo.generate_report(df=df.head(10), path_to_save='./Data_Profile_Report.html', pdf=False)
    schema = wombatoo.schema(df=df,path_to_save='schema.csv',show=1)

END