topper 1.0.1

An application computing top songs by country or user_id over 7 days

Topper

Topper is a library made to parse and process log files of music listening. The purpose of this application is to get the top 50 songs the most listened on the last 7 days grouped by country or user id.

You must provide:

• landing_folder: where daily files are sent
• checkpoint_directory: used by the application to process, archive and persist data across days
• output_directory: where result files are written everyday
• mode (optional): country (default) or user is the aggregation mode

Input files

Each log file must match the pattern listen-YYYYMMDD.log. The file must contains data structured with:

• One row per stream (1 listening).
• Each row is in the following format: song_id|user_id|country

File management

Input files will be moved to the directory checkpoint/current/

Invalid files will be moved to the directory checkpoint/errors/

Data files older than 7 days will be moved to the directory checkpoint/archive/

Output files

Mode 'country'

Produced files have the following format:

country1|sng_id1:n1,sng_id2:n2,sng_id3:n3,...,sng_id50:n50
country2|sng_id1:n1,sng_id2:n2,sng_id3:n3,...,sng_id50:n50

Where country is the country ISO2 code, sng_id1:n1 the identifier of the song the most listened with the related number of streams, sng_id2:n2 the identifier of the 2nd song the most listened with the related number of streams and so on..

Mode 'user'

Produced files have the following format:

user_id1|sng_id1:n1,sng_id2:n2,sng_id3:n3,...,sng_id50:n50
user_id2|sng_id1:n1,sng_id2:n2,sng_id3:n3,...,sng_id50:n50

Where user_id is a user, sng_id1:n1 the identifier of the song the most listened with the related number of streams, sng_id2:n2 the identifier of the 2nd song the most listened with the related number of streams and so on..

Usage

Supported Python Versions

Python >= 3.6

Installation

virtualenv -p python3 venv
source venv/bin/activate
make install

Display usage

topper -h

Example

topper --landing_folder sample/ --checkpoint_directory checkpoint --output_directory output --mode country

Development

make test # coverage tests
make linter # runs pylint
make build