PY-RETROSHEET

Python scripts for Retrosheet data downloading and parsing.

YE REQUIREMENTS

• Chadwick 0.6.2 http://chadwick.sourceforge.net/

• python 2.5+ (don't know about 3.0, sorry)

• sqlalchemy: http://www.sqlalchemy.org/

• [if using postgres] pyscopg2 python package (dependency for sqlalchemy)

USAGE

Download

python download.py [-y <4-digit-year> | --year <4-digit-year>]

The scripts/download.py script downloads Retrosheet data. Edit the config.ini file to configure what types of files should be downloaded. Optionally set the year to download via the command line argument.

• download > dl_eventfiles determines if Retrosheet Event Files should be downloaded or not. These are the only files that can be processed by parse.py at this time.

• download > dl_gamelogs determines if Retrosheet Game Logs should be downloaded or not. These are not able to be processed by parse.py at this time.

Parse into SQL

python parse.py [-y <4-digit-year>]

After the files have been downloaded, parse them into SQL with parse.py.

1. Create database called retrosheet (or whatever).

2. Add schema to the database w/ the included SQL script (the .postgres.sql one works nicely w/ PG, the other w/ MySQL)

3. Configure the file config.ini with your appropriate ENGINE, USER, HOST, PASSWORD, and DATABASE values - if you're using postgres, you can optionally define SCHEMA and download directory

   • Valid values for ENGINE are valid sqlalchemy engines e.g. 'mysql', 'postgresql', or 'sqlite',

   • If you have your server configured to allow passwordless connections, you don't need to define USER and PASSWORD.

   • If you are using sqlite3, database in the config should be the path to your database file.

   • Specify directory for retrosheet files to be downloaded to, needs to exist before script runs

4. Run parse.py to parse the files and insert the data into the database. (optionally use -y YYYY to import just one year)

YE GRATITUDE

Github user jeffcrow made many fixes and additions and added sqlite support

JUST THE DATA

If you're using PostgreSQL (and you should be), you can get a dump of all data up through 2014 (warning: 502MB) here