mlb boxscores

mlb.py

#! /usr/bin/python
import urllib2 
from datetime import datetime, timedelta
import simplejson as JSON
from pprint import pprint
import re

when = datetime.now() #- timedelta(days=1)


GRID_URL_TEMPLATE = r'http://gd2.mlb.com/components/game/mlb/year_%(year)d/month_%(month)02d/day_%(day)02d/grid.json'
BOX_SCORE_URL_TEMPLATE = r'http://gd2.mlb.com/components/game/mlb/year_%(year)s/month_%(month)s/day_%(day)s/gid_%(year)s_%(month)s_%(day)s_%(teama)s_%(teamb)s_%(seq)s/boxscore.json'

#lazy/poor mans xpath for json ( no support for indicies or anything cool like that )
def pull( data, key_list, default ):
    if len( key_list ) == 0:
        return default
    if len( key_list ) == 1:
        return data.get( key_list[0], default )
    else:
        return pull( data[key_list[0]], key_list[1:], default )

def pull_by_path( data, path, default ):
    return pull( data, path.split('/'), default )

grid = urllib2.urlopen( GRID_URL_TEMPLATE % { 'day': when.day, 'month':when.month, 'year': when.year } ).read()
gamesdata = JSON.loads( grid )

for game in pull_by_path( gamesdata, 'data/games/game', [] ):
    print
    print "%s at %s -- %s" % ( game['away_team_name'], game['home_team_name'], game['event_time'] )
    m = re.match( r"(?P<year>\d+)/(?P<month>\d+)/(?P<day>\d+)/(?P<teama>\w+)-(?P<teamb>\w+)-(?P<seq>\d+)", game['id'] )

    try:
        boxscore = urllib2.urlopen( BOX_SCORE_URL_TEMPLATE % m.groupdict() ).read()
        boxscoredata = JSON.loads( boxscore )
        #print "box score:", boxscore
        #pprint( boxscoredata )
        print "home",
        print pull_by_path( boxscoredata, 'data/boxscore/home_sname', "NA"),
        print pull_by_path( boxscoredata, 'data/boxscore/home_wins', "NA"),
        print pull_by_path( boxscoredata, 'data/boxscore/home_loss', "NA")
        print "away",
        print pull_by_path( boxscoredata, 'data/boxscore/away_sname', "NA"),
        print pull_by_path( boxscoredata, 'data/boxscore/away_wins', "NA"),
        print pull_by_path( boxscoredata, 'data/boxscore/away_loss', "NA")


        print "boxscores"
        for inning in pull_by_path( boxscoredata, 'data/boxscore/linescore/inning_line_score', ['bogus',3,3,] ):
            print "%2s" % inning['inning'],
        print "%17s" % "team", 
        print "%7s" % "runs", 
        print "%7s" % "hits", 
        print "%7s" % "errors"

        for inning in pull_by_path( boxscoredata, 'data/boxscore/linescore/inning_line_score', ['bogus',3,3,] ):
            print "%2s" % inning['away'],
        print "%17s" % pull_by_path( boxscoredata,'data/boxscore/away_sname', "_"),
        print "%7s" % pull_by_path( boxscoredata,'data/boxscore/linescore/away_team_runs', "_"), 
        print "%7s" % pull_by_path( boxscoredata,'data/boxscore/linescore/away_team_hits', "_"), 
        print "%7s" % pull_by_path( boxscoredata,'data/boxscore/linescore/away_team_errors', "_")
        print

        for inning in pull_by_path( boxscoredata, 'data/boxscore/linescore/inning_line_score', ['bogus',3,3,] ):
            print "%2s" % inning['home'],
        print "%17s" % pull_by_path( boxscoredata,'data/boxscore/home_sname','_') ,
        print "%7s" % pull_by_path( boxscoredata, 'data/boxscore/linescore/home_team_runs','_') , 
        print "%7s" % pull_by_path( boxscoredata, 'data/boxscore/linescore/home_team_hits','_') , 
        print "%7s" % pull_by_path( boxscoredata, 'data/boxscore/linescore/home_team_errors','_' )
        print
    except Exception, e: #urllib2.HttpException, e:
        #print e
        pass

    print