etl.py

import os
import glob
import psycopg2
import pandas as pd
from sql_queries import *


def process_song_file(cur, filepath):
    """Read song file, insert song and artist data into the database.
    
    Song record inserts data for song_id, title, artist_id, year and duration.
    Artist record inserts data for artist_id, name, location, lattitude, longitude.
    """
    # open song file
    df = pd.read_json(filepath, lines=True)
    
    # change year datatype from numpy int64 to string
    # error occurs on database write with int64 data type
    df['year'] = df['year'].apply(lambda x: str(x))

    # insert song record
    song_data = df.iloc[0, [7, 8, 0, 9, 5]].values.tolist()
    cur.execute(song_table_insert, song_data)
    
    # insert artist record
    artist_data = df.iloc[0, [0, 4, 2, 1, 3]].values.tolist()
    cur.execute(artist_table_insert, artist_data)


def process_log_file(cur, filepath):
    """Read log file, process and load data into user and songplay tables.
    """
    # open log file
    df = pd.read_json(filepath,lines=True) 

    # filter by NextSong action
    df = df[df.page == 'NextSong']

    # convert timestamp column to datetime
    t = pd.to_datetime(df['ts'], unit='ms')
    df['ts'] = pd.to_datetime(df['ts'], unit='ms')
    
    # insert time data records
    time_data = [t,t.dt.hour,t.dt.day,t.dt.week,t.dt.month,t.dt.year,t.dt.weekday]
    column_labels = ("timestamp","hour","day","week","month","year","weekday") 
    
    time_dict = {}

    for i in range(0,len(column_labels)):
        time_dict.update({column_labels[i]:time_data[i]})
        
    time_df = pd.DataFrame.from_dict(time_dict)
    
    for i, row in time_df.iterrows():
        cur.execute(time_table_insert, list(row))

    # load user table
    user_df = df.iloc[:,[17,2,5,3,7]]

    # insert user records
    for i, row in user_df.iterrows():
        cur.execute(user_table_insert, row)

    # insert songplay records
    for index, row in df.iterrows():
        
        # get songid and artistid from song and artist tables
        results = cur.execute(song_select, (row.song, row.artist, row.length))
        results = cur.fetchone()
        
        if results:
            songid, artistid = results
        else:
            songid, artistid = None, None

        # insert songplay record
        songplay_data = list([row.ts, row.userId,row.level,songid,artistid
                              ,row.sessionId,row.location,row.userAgent])
        cur.execute(songplay_table_insert, songplay_data)


def process_data(cur, conn, filepath, func):
    """Identify and process datafiles, based on filepath and function passed
    """
    # get all files matching extension from directory
    all_files = []
    for root, dirs, files in os.walk(filepath):
        files = glob.glob(os.path.join(root,'*.json'))
        for f in files :
            all_files.append(os.path.abspath(f))


    # get total number of files found
    num_files = len(all_files)
    print('{} files found in {}'.format(num_files, filepath))

    # iterate over files and process
    for i, datafile in enumerate(all_files, 1):
        print('{} file processing.'.format(datafile))
        func(cur, datafile)
        conn.commit()
        print('{}/{} files processed.'.format(i, num_files))


def main():
    """Establish database connection, process song and log data.
    """
    conn = psycopg2.connect("host=127.0.0.1 dbname=sparkifydb user=student password=student")
    cur = conn.cursor()

    process_data(cur, conn, filepath='data/song_data', func=process_song_file)
    process_data(cur, conn, filepath='data/log_data', func=process_log_file)

    conn.close()


if __name__ == "__main__":
    main()