class NBAComParser_test(unittest.TestCase): def setUp(self): logging.getLogger(__name__).addHandler(logging.NullHandler()) self.nbs = NBAComScraper() self.nbp = NBAComParser() def _get_from_file(self, fn): # content is none if file does not exist content = None # test if file exists, if so, slurp it into content if os.path.isfile(fn): try: with open(fn) as x: content = x.read() except: logging.exception('could not read from file ' + fn) return content # Can try to access from file instead of using scraper def _player_gamelogs(self,fn='player_gamelogs.json'): return self._get_from_file(fn) def _player_info(self,fn='player_info.json'): return self._get_from_file(fn) def _players(self,fn='players.json'): return self._get_from_file(fn) def _playerstats(self,fn='player_stats.json'): return self._get_from_file(fn) def _scoreboard(self,fn='scoreboard.json'): return self._get_from_file(fn) def _team_gamelogs(self,fn='team_game_logs.json'): return self._get_from_file(fn) # START tests def test_boxscore(self): content = self.nbs.boxscore('0021500001', '2015-16') players, teams, starter_bench = self.nbp.boxscore(content) self.assertIsInstance(players, list) self.assertIsInstance(teams, list) self.assertIsInstance(players[0], dict) self.assertIn('MIN_PLAYED', players[0], "players should have min_played") self.assertIn('MIN_PLAYED', players[8], "players should have min_played") self.assertIn('MIN', teams[0], "teams should have min") self.assertIn('MIN', teams[1], "teams should have min") def test_one_player_gamelogs(self): content = self.nbs.one_player_gamelogs('203083', '2015-16') gls = self.nbp.one_player_gamelogs(content) self.assertIsInstance(gls, list) self.assertIsInstance(gls[0], dict) def test_one_team_gamelogs(self): team_id = '1610612765' season = '2015-16' content = self.nbs.one_team_gamelogs(team_id, season) gls = self.nbp.one_team_gamelogs(content) self.assertIsInstance(gls, list) self.assertIsInstance(gls[0], dict) def test_player_info(self): content = self.nbs.player_info('201939', '2015-16') pinfo = self.nbp.player_info(content) self.assertIsInstance(pinfo, dict) def test_players(self): content = self.nbs.players(season='2015-16', IsOnlyCurrentSeason='1') ps = self.nbp.players(content) self.assertIsInstance(ps, list) self.assertIsNotNone(ps[0], 'ps should not be none') def test_playerstats(self): statdate = '2016-01-20' content = self.nbs.playerstats('2015-16') ps = self.nbp.playerstats(content, statdate) self.assertIsInstance(ps, list) self.assertEqual(ps[0].get('STATDATE'), statdate) def test_season_gamelogs(self): team_content = self.nbs.season_gamelogs('2015-16', 'T') player_content = self.nbs.season_gamelogs('2015-16', 'P') team_gl = self.nbp.season_gamelogs(team_content, 'T') player_gl = self.nbp.season_gamelogs(player_content, 'P') self.assertIsInstance(player_gl, list) self.assertIsInstance(team_gl, list) self.assertIn('GAME_ID', player_gl[0], "players should have game_id") self.assertIn('GAME_ID', team_gl[0], "teams should have game_id") def test_team_dashboard(self): #dashboard = {'parameters': [], 'overall': [], 'location': [], 'days_rest': [], 'wins_losses': []} team_id = '1610612765' season = '2015-16' content = self.nbs.team_dashboard(team_id, season) td = self.nbp.team_dashboard(content) self.assertIsInstance(td, dict) self.assertIn('overall', td, "dashboard has overall") def test_team_opponent_dashboard(self): season = '2015-16' content = self.nbs.team_opponent_dashboard(season) td = self.nbp.team_opponent_dashboard(content) self.assertIsInstance(td, list) self.assertIsNotNone(td[0], 'td should not be None') def test_teams(self): content = self.nbs.teams() tms = self.nbp.teams(content) self.assertIsInstance(tms, dict) self.assertIsNotNone(tms, 'tms should not be None') def test_teamstats(self): season = '2015-16' content = self.nbs.teamstats(season) ts = self.nbp.teamstats(content) self.assertIsInstance(ts, list) self.assertIsNotNone(ts[0], 'ts should not be None')
class NBAComAgent(object): ''' Performs script-like tasks using NBA.com API ''' def __init__(self, db=None, cache_name=None, cookies=None): ''' Arguments: cache_name: str for scraper cache_name cookies: cookie jar db: NBAComPg instance ''' logging.getLogger(__name__).addHandler(logging.NullHandler()) self.scraper = NBAComScraper(cache_name=cache_name, cookies=cookies) self.parser = NBAComParser() if db: self.db = db self.insert_db = True else: self.insert_db = False def _combined_player_boxscores(self, gid): ''' Combines 5 types of boxscores from nba.com API into list of boxscores Arguments: gid: string game ID, with leading '00' Returns: list of player boxscores combined - traditional, advanced, misc, scoring, usage Examples: a = NBAComAgent() playerboxes = a.combined_player_boxscores('0020161001') ''' traditional_players, traditional_teams, traditional_starter_bench = self.parser.boxscore_traditional( self.scraper.boxscore_traditional(gid)) adv_players, adv_team = self.parser.boxscore_advanced( self.scraper.boxscore_advanced(gid)) misc_players, misc_team = self.parser.boxscore_misc( self.scraper.boxscore_misc(gid)) scoring_players, scoring_team = self.parser.boxscore_scoring( self.scraper.boxscore_scoring(gid)) usage_players = self.parser.boxscore_usage( self.scraper.boxscore_usage(gid)) # now need to combine player and team boxscores players = merge( dict(), [{t['PLAYER_ID']: t for t in traditional_players}, {t['PLAYER_ID']: t for t in adv_players}, {t['PLAYER_ID']: t for t in misc_players}, {t['PLAYER_ID']: t for t in scoring_players}, {t['PLAYER_ID']: t for t in usage_players}]) return players.values() def combined_team_boxscores(self, gid): ''' Combines 5 types of boxscores from nba.com API into list of boxscores Arguments: gid: string game ID, with leading '00' Returns: list of team boxscores - combined traditional, advanced, misc, scoring Examples: a = NBAComAgent() teamboxes = a.combined_team_boxscores('0020161001') ''' traditional_players, traditional_teams, traditional_starter_bench = self.parser.boxscore_traditional( self.scraper.boxscore_traditional(gid)) adv_players, adv_teams = self.parser.boxscore_advanced( self.scraper.boxscore_advanced(gid)) misc_players, misc_teams = self.parser.boxscore_misc( self.scraper.boxscore_misc(gid)) scoring_players, scoring_teams = self.parser.boxscore_scoring( self.scraper.boxscore_scoring(gid)) # now need to combine player and team boxscores teams = merge( dict(), [{t['TEAM_ID']: t for t in traditional_teams}, {t['TEAM_ID']: t for t in adv_teams}, {t['TEAM_ID']: t for t in misc_teams}, {t['TEAM_ID']: t for t in scoring_teams}]) return list(teams.values()) def linescores(self): ''' Updates gamesmeta table with game_information ''' #q = """SELECT '00' || game_id, to_char(game_date, 'YYYYmmdd') FROM gamesmeta # WHERE season = (select max(season) from seasons) AND game_date < now()::date AND q1 IS NULL # ORDER BY game_date DESC;""" #q = """SELECT '00' || game_id as gid, to_char(game_date, 'YYYYmmdd') as gd FROM cs_games # WHERE game_date < (CURRENT_TIMESTAMP AT TIME ZONE 'CST')::date AND # game_id NOT IN (SELECT DISTINCT game_id FROM boxv2015) # ORDER BY game_date DESC;""" q = """SELECT '00' || game_id as gid, to_char(game_date, 'YYYYmmdd') as gd FROM games WHERE game_date < localdate() AND season > 2015 AND game_id NOT IN (SELECT DISTINCT game_id FROM boxv2015) ORDER BY game_date DESC;""" for g in self.db.select_dict(q): try: content = self.scraper.boxscore_v2015(g['gid'], g['gd']) v, h = self.parser.boxscore_v2015(content) self.db.insert_dicts([v, h], 'boxv2015') logging.info('finished {} - {}'.format(g['gd'], g['gid'])) except Exception as e: logging.error('could not get {}'.format(g)) logging.exception(e) def new_players(self, season): ''' Updates players table with missing players Arguments: season (str): in YYYY-YY format Returns: list of players to add to stats.players Examples: a = NBAComAgent(cache_name='newplayers', cookies=httplib.CookieJar(), db=NBAComPg(...)) np = a.new_players(season='2015-16') ''' content = self.scraper.players_v2015(season) players = self.parser.players_v2015(content) currids = set([int(p.get('personId', 0)) for p in players]) logging.debug(currids) allids = set( self.db.select_list('SELECT nbacom_player_id from players')) missing = currids - allids if missing: np = [p for p in players if int(p['personId']) in missing] if self.insert_db: self.db.insert_players_v2015(np) return np else: return None def player_boxscores_combined(self): ''' Fetches player boxscores combined Arguments: season: str in YYYY-YY format (2015-16) Returns: players (list): player boxscores ''' pboxes = [] gids = self.db.select_list(missing_player_boxscores()) if not gids: logging.error('no missing gameids found') return None logging.info('there are {} missing game boxscores'.format(len(gids))) for gid in gids: logging.info('getting {}'.format(gid)) box = self._combined_player_boxscores(gid) if not box: logging.error('no box for {}'.format(gid)) continue if self.insert_db: self.db.insert_player_boxscores(box) pboxes.append(box) return [item for sublist in pboxes for item in sublist] def player_boxscores_combined(self): ''' Fetches player boxscores combined Arguments: season: str in YYYY-YY format (2015-16) Returns: players (list): player boxscores ''' pboxes = [] gids = self.db.select_list(missing_player_boxscores()) if not gids: logging.error('no missing gameids found') return None logging.info('there are {} missing game boxscores'.format(len(gids))) for gid in gids: logging.info('getting {}'.format(gid)) box = self._combined_player_boxscores(gid) if not box: logging.error('no box for {}'.format(gid)) continue if self.insert_db: self.db.insert_player_boxscores(box) pboxes.append(box) return [item for sublist in pboxes for item in sublist] def player_gamelogs(self, season, date_from=None, date_to=None): ''' Fetches player_gamelogs and updates player_gamelogs table Arguments: season: str in YYYY-YY format (2015-16) date_from: str in YYYY-mm-dd format date_to: str in YYYY-mm-dd format Returns: players (list): player dictionary of stats + dfs points ''' pgl = self.parser.season_gamelogs( self.scraper.season_gamelogs(season, 'P'), 'P') if self.insert_db: self.db.insert_player_gamelogs(pgl) return pgl def playerstats(self, season, date_from=None, date_to=None, all_missing=False): ''' Fetches playerstats and updates player_gamelogs table Arguments: season: str in YYYY-YY format (2015-16) date_from: str in YYYY-mm-dd format date_to: str in YYYY-mm-dd format all_missing: boolean Returns: players (list): player dictionary of stats + dfs points Examples: a = NBAComAgent() np = a.playerstats(season='2015-16', date_from='2016-03-01', date_to='2016-03-08') ''' if date_from and date_to: ps_base = self.parser.playerstats( self.scraper.playerstats(season, DateFrom=date_from, DateTo=date_to)) ps_advanced = self.parser.playerstats( self.scraper.playerstats(season, DateFrom=date_from, DateTo=date_to, MeasureType='Advanced')) ps = [ merge(dict(), [psb, psadv]) for psb, psadv in zip(ps_base, ps_advanced) ] if self.insert_db: self.db.insert_playerstats(ps, as_of=date_to) return ps elif all_missing: pstats = {} start = datetostr(d=season_start(season), site='nba') for day in self.db.select_list(missing_playerstats()): daystr = datetostr(day, 'nba') ps_base = self.parser.playerstats( self.scraper.playerstats(season, DateFrom=start, DateTo=daystr)) ps_advanced = self.parser.playerstats( self.scraper.playerstats(season, DateFrom=start, DateTo=daystr, MeasureType='Advanced')) ps = [ merge(dict(), [psadv, psb]) for psb, psadv in zip(ps_base, ps_advanced) ] pstats[daystr] = ps if self.insert_db: self.db.insert_playerstats(ps, as_of=daystr) logging.info('completed {}'.format(daystr)) return pstats else: raise ValueError( 'need to specify dates or set all_missing to true') def scoreboards(self, season_start, season_end): ''' Downloads and parses range of scoreboards Arguments: season_start (str): in %Y-%m-%d format season_end (str): in %Y-%m-%d format Returns: scoreboards (list): scoreboard dicts Examples: a = NBAComAgent() sb = a.scoreboards(season_start='2015-10-27', season_end='2016-04-15') ''' scoreboards = [] for day in reversed(date_list(season_end, season_start)): game_date = dt.datetime.strftime(day, '%Y-%m-%d') scoreboard_json = self.scraper.scoreboard(game_date=game_date) scoreboard = self.parser.scoreboard(scoreboard_json, game_date=game_date) scoreboards.append(scoreboard) if self.insert_db: self.db.insert_scoreboards(scoreboards) return scoreboards def team_boxscores_combined(self): ''' Fetches team boxscores combined Returns: tboxes: list of boxscores ''' tboxes = [] gids = self.db.select_list(missing_team_boxscores()) if not gids: logging.error('no missing gameids found') return None logging.info('there are {} missing game boxscores'.format(len(gids))) for gid in gids: logging.info('getting {}'.format(gid)) box = self.combined_team_boxscores(gid) if not box: logging.error('no box for {}'.format(gid)) continue if self.insert_db: self.db.insert_team_boxscores(box) tboxes.append(box) return [item for sublist in tboxes for item in sublist] def team_gamelogs(self, season, date_from=None, date_to=None): ''' Fetches team_gamelogs and updates cs_team_gamelogs table Arguments: season (str): in YYYY-YY format (2015-16) Returns: team_gl (list): player dictionary of stats Examples: a = NBAComAgent() tgl = a.team_gamelogs(season='2015-16', date_from='2016-03-01', date_to='2016-03-08', insert_db=True) ''' content = self.scraper.season_gamelogs(season=season, player_or_team='T') tgl = self.parser.season_gamelogs(content, 'T') mtgl = self.db.missing_tgl() if tgl and mtgl: toins = [gl for gl in tgl if gl.get('GAME_ID', None) in mtgl] if self.insert_db: self.db.insert_team_gamelogs(toins) return toins else: logging.error('no team gamelogs to insert') def teamstats(self, season, date_from=None, date_to=None, all_missing=False): ''' Fetches teamstats and updates database table Arguments: season (str): in YYYY-YY format (2015-16) date_from (str): in %Y-%m-%d format, default beginning of season date_from (str): in %Y-%m-%d format, default yesterday all_missing: boolean Returns: list of team dictionary of basic and advanced stats Examples: a = NBAComAgent() ps = a.teamstats(season='2015-16', date_from='2016-03-01', date_to='2016-03-08') ps = a.teamstats(season='2015-16', all_missing=True) ''' if date_from and date_to: ts_base = self.parser.teamstats( self.scraper.teamstats(season, DateFrom=date_from, DateTo=date_to)) ts_advanced = self.parser.teamstats( self.scraper.teamstats(season, DateFrom=date_from, DateTo=date_to, MeasureType='Advanced')) ts = [ merge(dict(), [psb, psadv]) for psb, psadv in zip(ts_base, ts_advanced) ] if self.insert_db: self.db.insert_teamstats(ts, as_of=date_to) return ts elif all_missing: tstats = {} start = datetostr(d=season_start(season), site='nba') for day in self.db.select_list(missing_teamstats()): daystr = datetostr(day, 'nba') ts_base = self.parser.teamstats( self.scraper.teamstats(season, DateFrom=start, DateTo=daystr)) ts_advanced = self.parser.teamstats( self.scraper.teamstats(season, DateFrom=start, DateTo=daystr, MeasureType='Advanced')) ts = [ merge(dict(), [psb, psadv]) for psb, psadv in zip(ts_base, ts_advanced) ] tstats[daystr] = ts if self.insert_db: self.db.insert_teamstats(ts, as_of=daystr) logging.debug('teamstats: completed {}'.format(daystr)) else: logging.error('did not insert: {}'.format(ts)) return tstats else: raise ValueError( 'need to specify dates or set all_missing to true') def team_opponent_dashboards(self, season, date_from=None, date_to=None, all_missing=False): ''' Downloads and parses range of team_opponents Arguments: season (str): in YYYY-YY format date_from (str): in %Y-%m-%d format, default is actual start of season date_to (str): in %Y-%m-%d format, default is actual end of season all_missing (bool): get all missing dashboards Returns: topp (list): dicts Examples: a = NBAComAgent() topp = a.team_opponent_dashboards('2014-15') ''' if date_from and date_to: content = self.scraper.team_opponent_dashboard(season, DateFrom=date_from, DateTo=date_to) topp = self.parser.team_opponent_dashboard(content) if self.insert_db: self.db.insert_team_opponent_dashboards(topp, as_of=date_to) return topp elif all_missing: topps = {} start = datetostr(d=season_start(season), site='nba') for day in self.db.select_list(missing_team_opponent_dashboard()): daystr = datetostr(day, 'nba') content = self.scraper.team_opponent_dashboard(season, DateFrom=start, DateTo=daystr) topp = self.parser.team_opponent_dashboard(content) if self.insert_db: self.db.insert_team_opponent_dashboards(topp, as_of=daystr) topps[daystr] = topp return topps else: raise ValueError( 'need to specify dates or set all_missing to true') def update_player_positions(self): ''' Trying to make sure all position data is current Only info in nba.com is PLAYER key, this is only Guard, etc. Unclear where the PG, etc. comes from TODO: this is not functional yet ''' if not self.db: raise ValueError('need database connection to update players') q = """SELECT nbacom_player_id FROM stats.players2 WHERE nbacom_position IS NULL or nbacom_position = ''""" uq = """UPDATE stats.players2 SET nbacom_position = '{}' WHERE nbacom_player_id = {}""" for pid in self.db.select_list(q): logging.debug('getting {}'.format(pid)) pinfo = self.parser.player_info( self.scraper.player_info(pid, '2015-16')) if pinfo.get('POSITION'): self.db.update(uq.format(pinfo.get('POSITION'), pid)) logging.debug('inserted {}'.format( pinfo.get('DISPLAY_FIRST_LAST')))
class NBAPlayers(object): ''' TODO: need to update / merge this with the names module Provides for updating nba.com players table (stats.players) Also permits cross-reference of player names and player ids from various sites(stats.player_xref) Usage: np = NBAComPlayers(db=True) np.missing_players('2015-16') ''' def __init__(self, db=False): self.logger = logging.getLogger(__name__) self._dk_players = [] self.scraper = NBAComScraper() self.parser = NBAComParser() if db: self.nbadb = NBAPostgres() @property def dk_players(self): ''' Dictionary of player_name: draftkings_player_id Returns: dk_players (dict): key is player name, value is dk_player_id ''' if not self._dk_players: try: fn = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'data', 'dk_players.json') with open(fn, 'r') as infile: self._dk_players = json.load(infile) except: self.logger.exception('could not open dk_players json file') return self._dk_players def missing_players(self, season): ''' Looks for missing players by comparing current_season_gamelogs and players tables Fetches player_info by id, inserts player into players table Arguments: season (str): example '2015-16' for the current season Returns: missing (list): list of dictionaries that represent row in players table ''' if not self.nbadb: raise ValueError('missing_players requires a database connection') missing = [] # get list of ids that appear in current_season_gamelogs but not players sql = ''' SELECT DISTINCT player_id FROM stats.cs_player_gamelogs WHERE player_id NOT IN (SELECT DISTINCT nbacom_player_id FROM stats.players) ''' results = self.nbadb.select_list(sql) if results: for pid in results: content = self.scraper.player_info(pid, season) pi = self.parser.player_info(content) pi = {k.lower(): v for k,v in pi.iteritems()} pi.pop('games_played_flag', None) pi['nbacom_team_id'] = pi.get('team_id', None) pi.pop('team_id', None) pi['nbacom_position'] = pi.get('position', None) pi.pop('position', None) pi['nbacom_player_id'] = pi.get('person_id', None) pi.pop('person_id', None) if pi.get('height', None): try: feet, inches = pi['height'].split('-') pi['height'] = int(feet)*12 + int(inches) except: pass # have to convert empty strings to None, otherwise insert fails for integer/numeric columns player_info= {} for k,v in pi.iteritems(): if not v: player_info[k] = None else: player_info[k] = v missing.append(player_info) if self.polite: time.sleep(1) if missing: self.nbadb.insert_dicts(missing, 'stats.players') return missing def recent_nbacom_players(self): ''' Gets last couple of years of players from nba.com Returns dictionary with key of name + team ''' if not self.nbadb: return None else: # can also return in dict format # {'{0} {1}'.format(item['player_name'], item.get('team_code')):item for item in self.nbadb.select_dict(sql)} sql = '''SELECT * FROM recent_nba_players''' return self.nbadb.select_dict(sql) def player_xref(self, site_name): ''' Obtains dictionary of site_player_id and nbacom_player_id Args: site_name (str): 'dk', 'fantasylabs', etc. Return: player_xref (dict): key is site_player_id ''' sql = '''SELECT nbacom_player_id, site_player_id FROM stats.player_xref WHERE site='{0}' ORDER by nbacom_player_id''' xref = self.nbadb.select_dict(sql.format(site_name)) return {p.get('site_player_id'): p.get('nbacom_player_id') for p in xref if p.get('site_player_id')} def site_to_nbacom(self, site, player_name): ''' Returns dictionary with name of player on site, value is list of name and id of player on nba.com Arguments: site (str): 'dk', 'fd', 'rg', 'doug', 'espn' Returns: players (dict): key is player name on site, value is list [nbacom_player_name, nbacom_player id] ''' if site.lower() not in self.sites: # should try name matcher pass elif site.lower() == 'dk': return self._dk_name(player_name) elif 'doug' in site.lower(): return self._doug_name(player_name) elif 'espn' in site.lower(): return self._espn_name(player_name) elif site.lower() == 'fd' or 'duel' in site.lower(): return self._fd_name(player_name) elif site.lower() == 'fl' or 'labs' in site.lower(): return self._fl_name(player_name) elif site.lower() == 'rg' or 'guru' in site.lower(): return self._rg_name(player_name)
class NBAPlayers(object): ''' TODO: need to update / merge this with the names module Provides for updating nba.com players table (stats.players) Also permits cross-reference of player names and player ids from various sites(stats.player_xref) Usage: np = NBAComPlayers(db=True) np.missing_players('2015-16') ''' def __init__(self, db=None): logging.getLogger(__name__).addHandler(logging.NullHandler()) self._dk_players = [] self.scraper = NBAComScraper() self.parser = NBAComParser() if db: self.nbadb = db def _convert_heigt(self, h): ''' Converts height from 6-11 (feet-inches) to 73 (inches) Args: h(str): e.g. 6-7, 5-11, 7-1 Returns: height(int): e.g. 79, 71, 85 ''' try: f, i = h.split('-') return int(f) * 12 + int(i) except: return None @property def dk_players(self): ''' Dictionary of player_name: draftkings_player_id Returns: dk_players (dict): key is player name, value is dk_player_id ''' if not self._dk_players: try: fn = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'data', 'dk_players.json') with open(fn, 'r') as infile: self._dk_players = json.load(infile) except: logging.exception('could not open dk_players json file') return self._dk_players def missing_players(self, season): ''' Looks for missing players by comparing current_season_gamelogs and players tables Fetches player_info by id, inserts player into players table Arguments: season (str): example '2015-16' for the current season Returns: missing (list): list of dictionaries that represent row in players table ''' if not self.nbadb: raise ValueError('missing_players requires a database connection') missing = [] # get list of ids that appear in current_season_gamelogs but not players sql = '''SELECT * FROM stats.players_to_add''' for pid in self.nbadb.select_list(sql): content = self.scraper.player_info(pid, season) pinfo = self.parser.player_info(content) pi = {k.lower(): v for k, v in pinfo.items()} pi.pop('games_played_flag', None) pi['nbacom_team_id'] = pi.get('team_id', None) pi.pop('team_id', None) pi['nbacom_position'] = pi.get('position', None) pi.pop('position', None) pi['nbacom_player_id'] = pi.get('person_id', None) pi.pop('person_id', None) if pi.get('height', None): pi['height'] = self._convert_heigt(pi['height']) # have to convert empty strings to None, otherwise insert fails for integer/numeric columns player_info = {} for k, v in pi.items(): if not v: player_info[k] = None else: player_info[k] = v missing.append(player_info) if self.polite: time.sleep(1) if missing: self.nbadb.insert_dicts(missing, 'stats.players') return missing def player_xref(self, site_name): ''' Obtains dictionary of site_player_id and nbacom_player_id Args: site_name (str): 'dk', 'fantasylabs', etc. Return: player_xref (dict): key is site_player_id ''' sql = '''SELECT nbacom_player_id, site_player_id FROM stats.player_xref WHERE site='{0}' ORDER by nbacom_player_id''' xref = self.nbadb.select_dict(sql.format(site_name)) return { p.get('site_player_id'): p.get('nbacom_player_id') for p in xref if p.get('site_player_id') } def site_to_nbacom(self, site, player_name): ''' Returns dictionary with name of player on site, value is list of name and id of player on nba.com Arguments: site (str): 'dk', 'fd', 'rg', 'doug', 'espn' Returns: players (dict): key is player name on site, value is list [nbacom_player_name, nbacom_player id] ''' if site.lower() not in self.sites: # should try name matcher pass elif site.lower() == 'dk': return self._dk_name(player_name) elif 'doug' in site.lower(): return self._doug_name(player_name) elif 'espn' in site.lower(): return self._espn_name(player_name) elif site.lower() == 'fd' or 'duel' in site.lower(): return self._fd_name(player_name) elif site.lower() == 'fl' or 'labs' in site.lower(): return self._fl_name(player_name) elif site.lower() == 'rg' or 'guru' in site.lower(): return self._rg_name(player_name)
class NBAComParser_test(unittest.TestCase): def setUp(self): logging.getLogger(__name__).addHandler(logging.NullHandler()) self.nbs = NBAComScraper() self.nbp = NBAComParser() def _get_from_file(self, fn): if os.path.isfile(fn): try: with open(fn) as x: return x.read() except: logging.exception('could not read from file ' + fn) return None def _one_player_gamelogs(self, fn='one_player_gamelogs.json'): return self._get_from_file(fn) def _one_team_gamelogs(self, fn='one_team_gamelogs.json'): return self._get_from_file(fn) def _player_boxscore_traditional(self, fn='boxscore_traditional.json'): return self._get_from_file(fn) def _player_gamelogs(self, fn='player_gamelogs.json'): return self._get_from_file(fn) def _player_info(self, fn='player_info.json'): return self._get_from_file(fn) def _players(self, fn='players.json'): return self._get_from_file(fn) def _playerstats(self, fn='player_stats.json'): return self._get_from_file(fn) def _scoreboard(self, fn='scoreboard.json'): return self._get_from_file(fn) def _team_dashboard(self, fn='team_dashboard.json'): return self._get_from_file(fn) def _team_gamelogs(self, fn='team_game_logs.json'): return self._get_from_file(fn) def _team_opponent_dashboard(self, fn='team_opponent_dashboard.json'): return self._get_from_file(fn) def _teams(self, fn='teams.js'): return self._get_from_file(fn) def _teamstats(self, fn='teamstats.json'): return self._get_from_file(fn) def test_boxscore(self): content = self._player_boxscore_traditional() if not content: content = self.nbs.boxscore_traditional('0021500001') players, teams, starter_bench = self.nbp.boxscore_traditional(content) self.assertIsInstance(players, list) self.assertIsInstance(teams, list) self.assertIsInstance(players[0], dict) self.assertIn('MIN_PLAYED', players[0], "players should have min_played") self.assertIn('MIN_PLAYED', players[8], "players should have min_played") self.assertIn('MIN', teams[0], "teams should have min") self.assertIn('MIN', teams[1], "teams should have min") def test_one_player_gamelogs(self): content = self._one_player_gamelogs() if not content: content = self.nbs.one_player_gamelogs('203083', '2015-16') gls = self.nbp.one_player_gamelogs(content) self.assertIsInstance(gls, list) self.assertIsInstance(gls[0], dict) logging.info(gls) def test_one_team_gamelogs(self): content = self._one_team_gamelogs() if not content: team_id = '1610612765' season = '2015-16' content = self.nbs.one_team_gamelogs(team_id, season) gls = self.nbp.one_team_gamelogs(content) self.assertIsInstance(gls, list) self.assertIsInstance(gls[0], dict) def test_player_info(self): content = self._player_info() if not content: content = self.nbs.player_info('201939', '2015-16') pinfo = self.nbp.player_info(content) self.assertIsInstance(pinfo, dict) def test_players(self): content = self._players() if not content: content = self.nbs.players(season='2016-17', cs_only='1') ps = self.nbp.players(content) self.assertIsInstance(ps, list) self.assertIsNotNone(ps[0], 'ps should not be none') def test_playerstats(self): content = self._playerstats() if not content: statdate = '2016-01-20' content = self.nbs.playerstats('2015-16') ps = self.nbp.playerstats(content, statdate) self.assertIsInstance(ps, list) self.assertEqual(ps[0].get('STATDATE'), statdate) def test_season_gamelogs(self): team_content = self._team_gamelogs() player_content = self._player_gamelogs() if not team_content: team_content = self.nbs.season_gamelogs('2015-16', 'T') if not player_content: player_content = self.nbs.season_gamelogs('2015-16', 'P') team_gl = self.nbp.season_gamelogs(team_content, 'T') player_gl = self.nbp.season_gamelogs(player_content, 'P') self.assertIsInstance(player_gl, list) self.assertIsInstance(team_gl, list) self.assertIn('GAME_ID', player_gl[0], "players should have game_id") self.assertIn('GAME_ID', team_gl[0], "teams should have game_id") def test_team_dashboard(self): content = self._team_dashboard() if not content: team_id = '1610612765' season = '2015-16' content = self.nbs.team_dashboard(team_id, season) td = self.nbp.team_dashboard(content) self.assertIsInstance(td, dict) self.assertIn('overall', td, "dashboard has overall") def test_team_opponent_dashboard(self): content = self._team_opponent_dashboard() if not content: season = '2015-16' content = self.nbs.team_opponent_dashboard(season) td = self.nbp.team_opponent_dashboard(content) self.assertIsInstance(td, list) self.assertIsNotNone(td[0], 'td should not be None') def test_teams(self): content = self._teams() if not content: content = self.nbs.teams() tms = self.nbp.teams(content) self.assertIsInstance(tms, dict) self.assertIsNotNone(tms, 'tms should not be None') def test_teamstats(self): content = self._teamstats() if not content: season = '2015-16' content = self.nbs.teamstats(season) ts = self.nbp.teamstats(content) self.assertIsInstance(ts, list) self.assertIsNotNone(ts[0], 'ts should not be None')
class NBAComAgent(object): ''' Performs script-like tasks using NBA.com API ''' def __init__(self, db=None, cache_name=None, cookies=None): ''' Arguments: cache_name: str for scraper cache_name cookies: cookie jar db: NBAComPg instance ''' self.scraper = NBAComScraper(cache_name=cache_name, cookies=cookies) self.parser = NBAComParser() if db: self.db = db self.insert_db = True else: self.insert_db = False def _merge(self, merge_dico, dico_list): ''' See http://stackoverflow.com/questions/28838291/merging-multiple-dictionaries-in-python Args: merge_dico: dico_list: Returns: merged dictionary ''' for dico in dico_list: for key, value in dico.items(): if type(value) == type(dict()): merge_dico.setdefault(key, dict()) self._merge(merge_dico[key], [value]) else: merge_dico[key] = value return merge_dico def combined_player_boxscores(self, gid): ''' Combines 5 types of boxscores from nba.com API into list of boxscores Arguments: gid: string game ID, with leading '00' Returns: list of player boxscores combined - traditional, advanced, misc, scoring, usage Examples: a = NBAComAgent() playerboxes = a.combined_player_boxscores('0020161001') ''' traditional_players, traditional_teams, traditional_starter_bench = self.parser.boxscore_traditional( self.scraper.boxscore_traditional(gid)) adv_players, adv_team = self.parser.boxscore_advanced( self.scraper.boxscore_advanced(gid)) misc_players, misc_team = self.parser.boxscore_misc( self.scraper.boxscore_misc(gid)) scoring_players, scoring_team = self.parser.boxscore_scoring( self.scraper.boxscore_scoring(gid)) usage_players = self.parser.boxscore_usage( self.scraper.boxscore_usage(gid)) # now need to combine player and team boxscores players = self._merge( dict(), [{t['PLAYER_ID']: t for t in traditional_players}, {t['PLAYER_ID']: t for t in adv_players}, {t['PLAYER_ID']: t for t in misc_players}, {t['PLAYER_ID']: t for t in scoring_players}, {t['PLAYER_ID']: t for t in usage_players}]) return list(players.values()) def combined_team_boxscores(self, gid): ''' Combines 5 types of boxscores from nba.com API into list of boxscores Arguments: gid: string game ID, with leading '00' Returns: list of team boxscores - combined traditional, advanced, misc, scoring Examples: a = NBAComAgent() teamboxes = a.combined_team_boxscores('0020161001') ''' traditional_players, traditional_teams, traditional_starter_bench = self.parser.boxscore_traditional( self.scraper.boxscore_traditional(gid)) adv_players, adv_teams = self.parser.boxscore_advanced( self.scraper.boxscore_advanced(gid)) misc_players, misc_teams = self.parser.boxscore_misc( self.scraper.boxscore_misc(gid)) scoring_players, scoring_teams = self.parser.boxscore_scoring( self.scraper.boxscore_scoring(gid)) # now need to combine player and team boxscores teams = self._merge( dict(), [{t['TEAM_ID']: t for t in traditional_teams}, {t['TEAM_ID']: t for t in adv_teams}, {t['TEAM_ID']: t for t in misc_teams}, {t['TEAM_ID']: t for t in scoring_teams}]) return list(teams.values()) def new_players(self, season): ''' Updates players table with missing players Arguments: season (str): in YYYY-YY format Returns: list of players to add to stats.players Examples: a = NBAComAgent(cache_name='newplayers', cookies=httplib.CookieJar(), db=NBAComPg(...)) np = a.new_players(season='2015-16') ''' content = self.scraper.players(season, cs_only=1) players = self.parser.players(content) currids = set([int(p.get('PERSON_ID', 0)) for p in players]) allids = set( self.db.select_list('SELECT nbacom_player_id from players')) missing = currids - allids if missing: np = [ self.parser.player_info(self.scraper.player_info(pid, season)) for pid in missing ] if self.insert_db: self.db.insert_players(np) return np else: return None def player_gamelogs(self, season, date_from=None, date_to=None): ''' Fetches player_gamelogs and updates player_gamelogs table Arguments: season: str in YYYY-YY format (2015-16) date_from: str in YYYY-mm-dd format date_to: str in YYYY-mm-dd format insert_db: add list to database Returns: players (list): player dictionary of stats + dfs points Examples: a = NBAComAgent() np = a.player_gamelogs(season='2015-16', date_from='2016-03-01', date_to='2016-03-08', insert_db=True) ''' # mpgl -> list of game_ids for which there are no player gamelogs # filter gamelogs to those from missing game_ids pgl = self.parser.season_gamelogs( self.scraper.season_gamelogs(season, 'P'), 'P') mpgl = self.db.missing_pgl() if pgl and mpgl: pgl = [gl for gl in pgl if gl.get('GAME_ID', None) in mpgl] # now make sure you have no new players currids = set([int(p.get('PERSON_ID', 0)) for p in pgl]) allids = set( self.db.select_list('SELECT nbacom_player_id from players')) if self.insert_db: self.db.insert_players([ self.parser.player_info( self.scraper.player_info(pid, season)) for pid in currids - allids ]) self.db.insert_player_gamelogs(pgl) return pgl else: return None def playerstats(self, season, date_from=None, date_to=None, all_missing=False): ''' Fetches playerstats and updates player_gamelogs table Arguments: season: str in YYYY-YY format (2015-16) date_from: str in YYYY-mm-dd format date_to: str in YYYY-mm-dd format all_missing: boolean Returns: players (list): player dictionary of stats + dfs points Examples: a = NBAComAgent() np = a.playerstats(season='2015-16', date_from='2016-03-01', date_to='2016-03-08') ''' if date_from and date_to: ps_base = self.parser.playerstats( self.scraper.playerstats(season, DateFrom=date_from, DateTo=date_to)) ps_advanced = self.parser.playerstats( self.scraper.playerstats(season, DateFrom=date_from, DateTo=date_to, MeasureType='Advanced')) ps = [ self._merge(dict(), [psb, psadv]) for psb, psadv in zip(ps_base, ps_advanced) ] if self.insert_db: self.db.insert_playerstats(ps, as_of=date_to) return ps elif all_missing: pstats = {} start = datetostr(d=season_start(season), site='nba') for day in self.db.select_list( 'SELECT game_date FROM missing_playerstats'): daystr = datetostr(day, 'nba') ps_base = self.parser.playerstats( self.scraper.playerstats(season, DateFrom=start, DateTo=daystr)) ps_advanced = self.parser.playerstats( self.scraper.playerstats(season, DateFrom=start, DateTo=daystr, MeasureType='Advanced')) ps = [ self._merge(dict(), [psb, psadv]) for psb, psadv in zip(ps_base, ps_advanced) ] pstats[daystr] = ps if self.insert_db: self.db.insert_playerstats(ps, as_of=daystr) return pstats else: raise ValueError( 'need to specify dates or set all_missing to true') def scoreboards(self, season_start, season_end): ''' Downloads and parses range of scoreboards Arguments: season_start (str): in %Y-%m-%d format season_end (str): in %Y-%m-%d format Returns: scoreboards (list): scoreboard dicts Examples: a = NBAComAgent() sb = a.scoreboards(season_start='2015-10-27', season_end='2016-04-15') ''' scoreboards = [] for day in reversed(date_list(season_end, season_start)): game_date = dt.datetime.strftime(day, '%Y-%m-%d') scoreboard_json = self.scraper.scoreboard(game_date=game_date) scoreboard = self.parser.scoreboard(scoreboard_json, game_date=game_date) scoreboards.append(scoreboard) if self.insert_db: self.db.insert_scoreboards(scoreboards) return scoreboards def team_gamelogs(self, season, date_from=None, date_to=None): ''' Fetches team_gamelogs and updates cs_team_gamelogs table Arguments: season (str): in YYYY-YY format (2015-16) Returns: team_gl (list): player dictionary of stats Examples: a = NBAComAgent() tgl = a.team_gamelogs(season='2015-16', date_from='2016-03-01', date_to='2016-03-08', insert_db=True) ''' content = self.scraper.season_gamelogs(season=season, player_or_team='T') tgl = self.parser.season_gamelogs(content, 'T') mtgl = self.db.missing_tgl() if tgl and mtgl: toins = [gl for gl in tgl if gl.get('GAME_ID', None) in mtgl] if self.insert_db: self.db.insert_team_gamelogs(toins) return toins else: logging.error('no team gamelogs to insert') def teamstats(self, season, date_from=None, date_to=None, all_missing=False): ''' Fetches teamstats and updates database table Arguments: season (str): in YYYY-YY format (2015-16) date_from (str): in %Y-%m-%d format, default beginning of season date_from (str): in %Y-%m-%d format, default yesterday all_missing: boolean Returns: list of team dictionary of basic and advanced stats Examples: a = NBAComAgent() ps = a.teamstats(season='2015-16', date_from='2016-03-01', date_to='2016-03-08') ps = a.teamstats(season='2015-16', all_missing=True) ''' if date_from and date_to: ts_base = self.parser.teamstats( self.scraper.teamstats(season, DateFrom=date_from, DateTo=date_to)) ts_advanced = self.parser.teamstats( self.scraper.teamstats(season, DateFrom=date_from, DateTo=date_to, MeasureType='Advanced')) ts = [ self._merge(dict(), [psb, psadv]) for psb, psadv in zip(ts_base, ts_advanced) ] if self.insert_db: self.db.insert_teamstats(ts, as_of=date_to) return ts elif all_missing: tstats = {} start = datetostr(d=season_start(season), site='nba') for day in self.db.select_list( 'SELECT game_date FROM missing_teamstats'): daystr = datetostr(day, 'nba') ts_base = self.parser.teamstats( self.scraper.teamstats(season, DateFrom=start, DateTo=daystr)) ts_advanced = self.parser.teamstats( self.scraper.teamstats(season, DateFrom=start, DateTo=daystr, MeasureType='Advanced')) ts = [ self._merge(dict(), [psb, psadv]) for psb, psadv in zip(ts_base, ts_advanced) ] tstats[daystr] = ts if self.insert_db: self.db.insert_teamstats(ts, as_of=daystr) logging.info('teamstats: completed {}'.format(daystr)) else: logging.error('did not insert: {}'.format(ts)) return tstats else: raise ValueError( 'need to specify dates or set all_missing to true') def team_opponent_dashboards(self, season, date_from=None, date_to=None, all_missing=False): ''' Downloads and parses range of team_opponents Arguments: season (str): in YYYY-YY format date_from (str): in %Y-%m-%d format, default is actual start of season date_to (str): in %Y-%m-%d format, default is actual end of season all_missing (bool): get all missing dashboards Returns: topp (list): dicts Examples: a = NBAComAgent() topp = a.team_opponent_dashboards('2014-15') ''' if date_from and date_to: content = self.scraper.team_opponent_dashboard(season, DateFrom=date_from, DateTo=date_to) topp = self.parser.team_opponent_dashboard(content) if self.insert_db: self.db.insert_team_opponent_dashboards(topp, as_of=date_to) return topp elif all_missing: topps = {} start = datetostr(d=season_start(season), site='nba') for day in self.db.select_list( 'SELECT * FROM missing_team_opponent_dashboard'): daystr = datetostr(day, 'nba') content = self.scraper.team_opponent_dashboard(season, DateFrom=start, DateTo=daystr) topp = self.parser.team_opponent_dashboard(content) if self.insert_db: self.db.insert_team_opponent_dashboards(topp, as_of=daystr) topps[daystr] = topp return topps else: raise ValueError( 'need to specify dates or set all_missing to true') def update_player_positions(self): ''' Trying to make sure all position data is current Only info in nba.com is PLAYER key, this is only Guard, etc. Unclear where the PG, etc. comes from TODO: this is not functional yet ''' if not self.db: raise ValueError('need database connection to update players') q = """SELECT nbacom_player_id FROM stats.players2 WHERE nbacom_position IS NULL or nbacom_position = ''""" uq = """UPDATE stats.players2 SET nbacom_position = '{}' WHERE nbacom_player_id = {}""" for pid in self.db.select_list(q): logging.info('getting {}'.format(pid)) pinfo = self.parser.player_info( self.scraper.player_info(pid, '2015-16')) if pinfo.get('POSITION'): self.db.update(uq.format(pinfo.get('POSITION'), pid)) logging.info('inserted {}'.format( pinfo.get('DISPLAY_FIRST_LAST')))