import pandas as pd
import seaborn as sns
import os
from timeit import default_timer as timer

measure_list = []
df = pd.read_csv('dummy_data.csv')

for i in range(10):
    start = timer()
    df.to_csv('dummy_data.csv', index=False)
    measure_list.append({'type':'csv', 'operation':'write', 'time': timer()-start})

    start = timer()
    df.to_parquet('dummy_data.parquet')
    measure_list.append({'type':'parquet', 'operation':'write', 'time': timer()-start})

    start = timer()
    df.to_pickle('dummy_data.pkl')
    measure_list.append({'type':'pickle', 'operation':'write', 'time': timer()-start})
    
    start = timer()
    df.to_feather('dummy_data.ftr')
    measure_list.append({'type':'feather', 'operation':'write', 'time': timer()-start})

for i in range(10):
    start = timer()
    _ = pd.read_csv('dummy_data.csv')
    measure_list.append({'type':'csv', 'operation':'read', 'time': timer()-start})

    start = timer()
    _ = pd.read_parquet('dummy_data.parquet')
    measure_list.append({'type':'parquet', 'operation':'read', 'time': timer()-start})

    start = timer()
    _ = pd.read_pickle('dummy_data.pkl')
    measure_list.append({'type':'pickle', 'operation':'read', 'time': timer()-start})
        
    start = timer()
    _ = pd.read_feather('dummy_data.ftr')
    measure_list.append({'type':'feather', 'operation':'read', 'time': timer()-start})

result_df = pd.DataFrame(measure_list)

sns.set(rc={'figure.figsize':(7, 5)}, font='Arial', font_scale=1.3)
ax = sns.barplot(x='type', y='time', hue='operation', data=result_df)
for p in ax.patches:
    height = p.get_height()
    ax.text(p.get_x()+p.get_width()/2., height + 0.7, '{:1.2f}'.format(height), ha="center") 

print(f'csv file size: {os.path.getsize("dummy_data.csv")>>20} MB')
print(f'feather file size: {os.path.getsize("dummy_data.ftr")>>20} MB')
print(f'parquet file size: {os.path.getsize("dummy_data.parquet")>>20} MB')
print(f'pickle file size: {os.path.getsize("dummy_data.pkl")>>20} MB')