"""
John Rachlin
DS 2000: Intro to Programming with Data


Date: Thu Oct 27 19:14:37 2022

File: reading.py
    
Description:
    
    
"""


import dataproc as dp

def main():
    data = dp.read_table('sample.csv', coltypes=[str,str,float])
    sales = dp.extract_column(data, 2)
    print(sales)
    
    datadict = dp.read_table_to_dict('sample.csv', coltypes=[str,str,float])
    sales2 = dp.extract_column(datadict, 'sales')
    print(sales2)
    
    
    
    #duration,start_day,start_day_name,start_station,end_station,bike_id
    #602,23,Friday,MIT Vassar St,Packard's Corner - Commonwealth Ave at Brighton Ave,8250
    
    trips = dp.read_table_to_dict('trips_small.csv', coltypes=[int, int, str, str, str, int])
    #for trip in trips:
    #    print(trip)
    
        
    # Now you don't need to keep track of column positions
    print(trips[0])
    print(trips[0]['start_station'])
    
    # Extract days of the week
    dow = dp.extract_column(trips, 'start_day_name')
    #print(dow)
    
    # Average trip duration
    durations = dp.extract_column(trips, 'duration')
    
    
    
    #print("Average trip duration: ", dp.avg(durations))
    
    # Trip duration histogram
    import matplotlib.pyplot as plt
    plt.hist(durations, bins=50)
    
    

if __name__ == '__main__':
    main()