import pandas as pd
import numpy as np
import matplotlib.pyplot as plt


# read in the csv file and put into dataframe
df = pd.read_csv('/Users/joconnor/Desktop/Me/performance_feb_6_19.csv')

# isolate all tests run week of Jan 28
week_1 = df['date'] < '2019-02-02'

# run groupby over each action within week_1
grouped_week_1 = df[week_1].groupby(['action'], as_index=False).duration.mean()

# give a name to timeframe of measurement
named_week_1 = grouped_week_1.rename(columns={'duration': 'Jan 28'})

# rinse and repeat, isolate and get the mean of all actions week of Feb 4
week_2 = df['date'].between('2019-02-02', '2019-02-09')
grouped_week_2 =  df[week_2].groupby('action', as_index=False).duration.mean()
named_week_2 = grouped_week_2.rename(columns={'duration': 'Feb 4'})

# combine each week into a single dataframe
combined_view = pd.merge(named_week_1, named_week_2, on='action', how='right')

# find difference to calculate change in performance times
combined_view['diff'] = combined_view['Jan 28'] - combined_view['Feb 4']
print(combined_view)

# export into a csv. this saves to the same folder.
combined_view.to_csv('file_name.csv', sep='\t', encoding='utf-8', index=False)