def get_espn_game_id(date, home_team, away_team):
"""
Scrapes the day's schedule and gets the id for the given game
Ex: http://www.espn.com/nhl/scoreboard?date=20161024
:param date: format-> YearMonthDay-> 20161024
:param home_team: home team
:param away_team: away team
:return: 9 digit game id
"""
url = 'http://www.espn.com/nhl/scoreboard?date={}'.format(
date.replace('-', ''))
response = shared.get_url(url)
# If can't get or not there return None
if not response:
raise Exception
game_ids = get_game_ids(response)
games = get_teams(response)
for i in range(len(games)):
if home_team in games[i] or away_team in games[i]:
return game_ids[i]
def get_shifts(game_id):
"""
Given a game_id it returns a DataFrame with the shifts for both teams
Ex: http://www.nhl.com/scores/htmlreports/20162017/TV020971.HTM
:param game_id: the game
:return: DataFrame with all shifts or None
"""
game_id = str(game_id)
home_url = 'http://www.nhl.com/scores/htmlreports/{}{}/TH{}.HTM'.format(game_id[:4], int(game_id[:4])+1, game_id[4:])
away_url = 'http://www.nhl.com/scores/htmlreports/{}{}/TV{}.HTM'.format(game_id[:4], int(game_id[:4])+1, game_id[4:])
home = shared.get_url(home_url)
time.sleep(1)
away = shared.get_url(away_url)
time.sleep(1)
return home, away
def get_roster(game_id):
"""
Given a game_id it returns the raw html
Ex: http://www.nhl.com/scores/htmlreports/20162017/RO020475.HTM
:param game_id: the game
:return: raw html of game
"""
game_id = str(game_id)
url = 'http://www.nhl.com/scores/htmlreports/{}{}/RO{}.HTM'.format(game_id[:4], int(game_id[:4]) + 1, game_id[4:])
return shared.get_url(url)
def get_schedule(date_from, date_to):
"""
Scrapes games in date range
Ex: http://statsapi.web.nhl.com/api/v1/schedule?startDate=2010-10-03&endDate=2011-06-20
:param date_from: scrape from this date
:param date_to: scrape until this date
:return: raw json of schedule of date range
"""
url = 'http://statsapi.web.nhl.com/api/v1/schedule?startDate={a}&endDate={b}'.format(a=date_from, b=date_to)
response = shared.get_url(url)
time.sleep(1)
return json.loads(response.text)
def get_shifts(game_id):
"""
Given a game_id it returns the raw json
Ex: http://www.nhl.com/stats/rest/shiftcharts?cayenneExp=gameId=2010020001
:param game_id: the game
:return: json or None
"""
url = 'http://www.nhl.com/stats/rest/shiftcharts?cayenneExp=gameId={}'.format(
game_id)
response = shared.get_url(url)
time.sleep(1)
# Return None if can't get page
if not response:
return None
return json.loads(response.text)
def get_espn(date, home_team, away_team):
"""
Gets the ESPN pbp feed
Ex: http://www.espn.com/nhl/gamecast/data/masterFeed?lang=en&isAll=true&gameId=400885300
:param date: date of the game
:param home_team: home team
:param away_team: away team
:return: raw xml
"""
game_id = get_espn_game_id(date, home_team.upper(), away_team.upper())
url = 'http://www.espn.com/nhl/gamecast/data/masterFeed?lang=en&isAll=true&gameId={}'.format(
game_id)
response = shared.get_url(url)
if response is None:
raise Exception
time.sleep(1)
return response
def get_pbp(game_id):
"""
Given a game_id it returns the raw json
Ex: http://statsapi.web.nhl.com/api/v1/game/2016020475/feed/live
:param game_id: the game
:return: raw json of game or None if couldn't get game
"""
url = 'http://statsapi.web.nhl.com/api/v1/game/{}/feed/live'.format(
game_id)
response = shared.get_url(url)
time.sleep(1)
# Return None if can't get page
if not response:
print("Json pbp for game {} is either not there or can't be obtained".
format(game_id))
return None
return json.loads(response.text)