def handle(self, *args, **options):
year = options['year']
# load MLB structure
call_command('load_MLB_structure', year)
# TODO load past players
# for now: load current rosters to at least have current players
call_command('load_players', year)
# load games from season
firstgame = mlbgame.important_dates(year).first_date_seas[:10]
lastgame = mlbgame.important_dates(year).last_date_seas[:10]
asg = mlbgame.important_dates(year).all_star_date[:10]
d = datetime.datetime.strptime(firstgame, '%Y-%m-%d')
while d < datetime.datetime.strptime(lastgame, '%Y-%m-%d'):
d += datetime.timedelta(1)
if d != datetime.datetime.strptime(asg, '%Y-%m-%d'):
call_command('load_day', d.year, d.month, d.day)
print('Complete')
def __fetch_important_dates(self):
today = datetime.today()
dates = mlbgame.important_dates()
if dates.playoffs_end_date != '':
season_end_date = datetime.strptime(dates.playoffs_end_date, MLB_DATE_FORMAT)
if (season_end_date - today).days < 0:
dates = mlbgame.important_dates(today.year + 1)
return dates
def crawl_xmls(self, browser):
xmls_to_fetch = [
"boxscore.xml", "rawboxscore.xml", "game_events.xml",
"linescore.xml", "players.xml", "inning/inning_all.xml", "game.xml"
]
event_dates = mlbgame.important_dates(year=self.year)
last_date = event_dates.last_date_seas
first_date = event_dates.first_date_seas if self.start_date is None else self.start_date
dates = date_range(first_date, last_date)
for date in dates:
game_scoreboards = mlbgame.day(date.year, date.month, date.day)
game_scoreboards = [[match] for match in game_scoreboards]
games = mlbgame.combine_games(game_scoreboards)
year = date.year
month = "{0:02d}".format(date.month)
day = "{0:02d}".format(date.day)
for game in games:
gameid = game.game_id
if not os.path.exists("xml/" + gameid):
os.makedirs("xml/" + gameid)
dir_path = "http://gd2.mlb.com/components/game/mlb/year_%s/month_%s/day_%s/gid_%s/" % \
(year, month, day, gameid)
for xml in xmls_to_fetch:
url_fetch = dir_path + xml
path_write = "xml/" + gameid + "/" + xml
self.fetch_xml(url_fetch, path_write)
def create_json(input_folder, input_summaries, output_folder):
for filename in os.listdir(input_folder):
d = None
with codecs.open(input_folder + filename) as json_data:
d = json.load(json_data)
print 'filename', input_folder + filename
output = []
for entry in d:
datetime_object = datetime.strptime(entry['day'], '%m_%d_%y')
begin_date = mlbgame.important_dates(
datetime_object.year).first_date_seas
begin_date = datetime.strptime(begin_date, '%Y-%m-%dT%H:%M:%S')
if datetime_object < begin_date:
print 'datetime_object', datetime_object, filename
continue
html_file_name = []
html_file_name.append(datetime_object.strftime("%Y%m%d"))
visname_homename = entry['vis_name'].replace(
" ", "_") + "-" + entry['home_name'].replace(" ", "_")
visname_homename = visname_homename.replace(
'D-backs', 'Diamondbacks')
html_file_name.append(visname_homename)
html_file_name.append(
str(entry['vis_line']['team_runs']) + "-" +
str(entry['home_line']['team_runs']))
files = glob.glob(input_summaries + "*" + "_".join(html_file_name))
if len(files) < 1:
print input_summaries + "*" + "_".join(
html_file_name) + " not found"
elif len(files) > 1:
print input_summaries + "*" + "_".join(
html_file_name) + " multiple found"
else:
fname = files[0]
with codecs.open(fname, encoding='utf-8') as f:
content = f.readlines()
# you may also want to remove whitespace characters like `\n` at the end of each line
content = [x.strip() for x in content]
text = " ".join(content)
words = nltk.word_tokenize(text)
newtokes = []
[
newtokes.append(toke) if toke[0].isupper()
or '-' not in toke else newtokes.extend(
toke.replace('-', " - ").split()) for toke in words
]
entry['summary'] = newtokes
output.append(entry)
if len(output) > 0:
with codecs.open(output_folder + 'combined_' + filename,
'w+') as outfile:
json.dump(output, outfile)
outfile.close()
def test_important_dates(self):
important_dates = mlbgame.important_dates(2017)
output = ("Opening Day 2017: Sunday, April 02.\n"
"Last day of the 1st half: Sunday, July 09.\n"
"2017 All Star Game: Tuesday, July 11.\n"
"First day of the 2nd half: Friday, July 14.\n"
"Last day of the 2017 season: Sunday, October 01.\n"
"2017 Playoffs start: Tuesday, October 03.\n"
"2017 Playoffs end: Wednesday, November 01.")
self.assertRaises(ValueError, lambda: mlbgame.important_dates(2050))
self.assertIsInstance(important_dates.organization_id, int)
self.assertIsInstance(important_dates.year, int)
self.assertIsInstance(important_dates.org_code, str)
self.assertIsInstance(important_dates.org_type, str)
self.assertIsInstance(important_dates.parent_org, str)
self.assertIsInstance(important_dates.parent_abbrev, str)
self.assertIsInstance(important_dates.name_full, str)
self.assertIsInstance(important_dates.name_short, str)
self.assertIsInstance(important_dates.name_abbrev, str)
self.assertIsInstance(important_dates.file_code, str)
self.assertIsInstance(important_dates.games, int)
self.assertIsInstance(important_dates.first_date_seas, str)
self.assertIsInstance(important_dates.last_date_1sth, str)
self.assertIsInstance(important_dates.first_date_2ndh, str)
self.assertIsInstance(important_dates.last_date_seas, str)
self.assertIsInstance(important_dates.split_season_sw, str)
self.assertIsInstance(important_dates.games_1sth, str)
self.assertIsInstance(important_dates.games_2ndh, str)
self.assertIsInstance(important_dates.all_star_sw, str)
self.assertIsInstance(important_dates.all_star_date, str)
self.assertIsInstance(important_dates.playoff_sw, str)
self.assertIsInstance(important_dates.playoff_teams, str)
self.assertIsInstance(important_dates.wildcard_sw, str)
self.assertIsInstance(important_dates.wildcard_teams, str)
self.assertIsInstance(important_dates.playoff_points_sw, str)
self.assertIsInstance(important_dates.point_values, str)
self.assertIsInstance(important_dates.playoffs_start_date, str)
self.assertIsInstance(important_dates.playoffs_end_date, str)
self.assertIsInstance(important_dates.playoff_rounds, str)
self.assertIsInstance(important_dates.playoff_games, str)
self.assertEqual(important_dates.organization_id, 1)
self.assertEqual(important_dates.year, 2017)
self.assertEqual(important_dates.org_code, 'mlb')
self.assertEqual(important_dates.org_type, 'S')
self.assertEqual(important_dates.parent_org, '')
self.assertEqual(important_dates.parent_abbrev, '')
self.assertEqual(important_dates.name_full, 'Major League Baseball')
self.assertEqual(important_dates.name_short, '')
self.assertEqual(important_dates.name_abbrev, 'MLB')
self.assertEqual(important_dates.file_code, 'mlb')
self.assertEqual(important_dates.games, 162)
self.assertEqual(important_dates.first_date_seas,
'2017-04-02T00:00:00')
self.assertEqual(important_dates.last_date_1sth, '2017-07-09T00:00:00')
self.assertEqual(important_dates.first_date_2ndh,
'2017-07-14T00:00:00')
self.assertEqual(important_dates.last_date_seas, '2017-10-01T00:00:00')
self.assertEqual(important_dates.split_season_sw, 'N')
self.assertEqual(important_dates.games_1sth, '')
self.assertEqual(important_dates.games_2ndh, '')
self.assertEqual(important_dates.all_star_sw, 'Y')
self.assertEqual(important_dates.all_star_date, '2017-07-11T00:00:00')
self.assertEqual(important_dates.playoff_sw, 'N')
self.assertEqual(important_dates.playoff_teams, '')
self.assertEqual(important_dates.wildcard_sw, 'N')
self.assertEqual(important_dates.wildcard_teams, '')
self.assertEqual(important_dates.playoff_points_sw, 'N')
self.assertEqual(important_dates.point_values, '')
self.assertEqual(important_dates.playoffs_start_date,
'2017-10-03T00:00:00')
self.assertEqual(important_dates.playoffs_end_date,
'2017-11-01T00:00:00')
self.assertEqual(important_dates.playoff_rounds, '')
self.assertEqual(important_dates.playoff_games, '')
self.assertEqual(mlbgame.info.date_format('2017-04-02T00:00:00'),
'Sunday, April 02')
self.assertEqual(mlbgame.info.date_format('not_a_date'), '')
self.assertEqual(important_dates.nice_output(), output)
self.assertEqual(important_dates.__str__(), output)
self.assertEqual(mlbgame.info.str_format('test-{0}', [1]), 'test-1')
mlbgame.data.IMPORTANT_DATES = '{0}'
self.assertRaises(ValueError,
lambda: mlbgame.data.get_important_dates(2050))
def search(request, team_searched):
fullNames = {
"Dodgers": "Los Angeles Dodgers",
"Indians": "Cleveland Indians",
"Rays": "Tampa Bay Rays",
"Twins": "Minnesota Twins",
"Athletics": "Oakland Athletics",
"White Sox": "Chicago White Sox",
"Reds": "Cincinnati Reds",
"Padres": "San Diego Padres",
"Cardinals": "St. Louis Cardinals",
"Cubs": "Chicago Cubs",
"Brewers": "Milwaukee Brewers",
"Royals": "Kansas City Royals",
"Astros": "Houston Astros",
"Yankees": "New York Yankees",
"Braves": "Atlanta Braves",
"Orioles": "Baltimore Orioles",
"Blue Jays": "Toronto Blue Jays",
"Giants": "San Francisco Giants",
"Pirates": "Pittsburgh Pirates",
"Diamondbacks": "Arizona Diamondbacks",
"Marlins": "Miami Marlins",
"Mets": "New York Mets",
"Mariners": "Seattle Mariners",
"Rangers": "Texas Rangers",
"Angels": "Los Angeles Angels",
"Nationals": "Washington Nationals",
"Phillies": "Philadelphia Phillies",
"Red Sox": "Boston Red Sox",
"Rockies": "Colorado Rockies",
"Tigers": "Detroit Tigers"
}
teamNames = {v: k for k, v in fullNames.items()}
searched = True
now = datetime.datetime.now()
areGamesToday = True
#retrieve today's games
games = mlbgame.day(now.year,
now.month,
now.day,
home=team_searched,
away=team_searched)[0:10]
dayCount = 1
#retrieve this week's games if not enough games today
while len(games) < 10 and dayCount < 7:
dateToCheck = datetime.date.today()
dateToCheck += datetime.timedelta(days=dayCount)
gamesOnDay = mlbgame.day(dateToCheck.year,
dateToCheck.month,
dateToCheck.day,
home=team_searched,
away=team_searched)[0:10 - len(games)]
if len(gamesOnDay) > 0:
games.append(gamesOnDay[0])
dayCount += 1
dayCount = 0
if len(games) == 0:
areGamesToday = False
#go to next season
date = mlbgame.important_dates(now.year + 1).first_date_seas.split('-')
year = int(date[0])
month = int(date[1])
day = int(date[2][:2])
while len(games) < 10 and dayCount < 6:
dateToCheck = datetime.date(year, month, day)
dateToCheck += datetime.timedelta(days=dayCount)
gamesOnDay = mlbgame.day(dateToCheck.year,
dateToCheck.month,
dateToCheck.day,
home=team_searched,
away=team_searched)[0:10]
if len(gamesOnDay) > 0:
for game in gamesOnDay:
games.append(game)
dayCount += 1
predictions = []
for game in games:
predictions.append(teamNames[predict(fullNames[game.home_team],
fullNames[game.home_team])])
zipped = list(zip(games, predictions))
return render(
request, 'home.html', {
'data': zipped,
'today': areGamesToday,
'team': team_searched,
'searched': searched
})
def set_mlb_dates(self):
if self.season:
self.mlb_dates = mlbgame.important_dates(self.season)
else:
self.mlb_dates = mlbgame.important_dates(self.current_season)