def generate_player(pos, row, args):
'''
Parses CSV row for DraftKings or FanDuel
and returns a player. Note that DraftKings
has different CSV formats for different
sports.
'''
avg_key = GAME_KEY_MAP[args.game]['avg']
name_key = GAME_KEY_MAP[args.game]['name']
team_key = GAME_KEY_MAP[args.game]['team']
team_alt_key = GAME_KEY_MAP[args.game]['team_alt']
game_key = GAME_KEY_MAP[args.game]['game']
game_alt_key = GAME_KEY_MAP[args.game]['game_alt']
avg = float(row.get(avg_key, 0))
player = Player(pos,
row[name_key],
row['Salary'],
possible_positions=row['Position'],
multi_position=('/' in row['Position']),
team=row.get(team_key) or row.get(team_alt_key),
matchup=row.get(game_key) or row.get(game_alt_key),
average_score=avg,
lock=(args.locked and row[name_key] in args.locked))
if args.source == 'DK_AVG':
player.proj = avg
return player
def generate_player(pos, row):
avg = float(row.get('AvgPointsPerGame', 0))
player = Player(pos,
row['Name'],
row['Salary'],
possible_positions=row['Position'],
multi_position=('/' in row['Position']),
team=row['teamAbbrev'],
matchup=row['GameInfo'],
average_score=avg,
lock=(args.locked and row['Name'] in args.locked))
if args.source == 'DK_AVG':
player.proj = avg
return player
def generate_player(pos, row):
# DK has inconsistent CSV formats
avg = float(row.get('AvgPointsPerGame', 0))
player = Player(
pos,
row['Name'],
row['Salary'],
possible_positions=row['Position'],
multi_position=('/' in row['Position']),
team=row.get('teamAbbrev') or row.get('TeamAbbrev'),
matchup=row.get('GameInfo') or row.get('Game Info'),
average_score=avg,
lock=(args.locked and row['Name'] in args.locked)
)
if args.source == _DK_AVG:
player.proj = avg
return player
def run(position_distribution, league, remove):
solver = pywraplp.Solver('FD',
pywraplp.Solver.CBC_MIXED_INTEGER_PROGRAMMING)
all_players = []
with open('data/dk-salaries-current-week.csv', 'rb') as csvfile:
csvdata = csv.DictReader(csvfile)
for idx, row in enumerate(csvdata):
if idx > 0:
player = Player(row['Position'], row['Name'], row['Salary'],
matchup=row['GameInfo'])
if args.l == 'NBA':
player.proj = float(row['AvgPointsPerGame'])
player.team = row['teamAbbrev']
all_players.append(player)
if league == 'NFL':
with open('data/fan-pros.csv', 'rb') as csvfile:
csvdata = csv.DictReader(csvfile)
mass_hold = [['playername', 'points', 'cost', 'ppd']]
for row in csvdata:
holder = row
player = filter(lambda x: x.name in row['playername'], all_players)
if len(player) == 0:
continue
player[0].proj = float(row['points'])
player[0].marked = 'Y'
player[0].team = row['playername'].split(' ')[-2]
listify_holder = [
row['playername'],
row['points']
]
if '0.0' not in row['points'] or player[0].cost != 0:
ppd = float(row['points']) / float(player[0].cost)
else:
ppd = 0
listify_holder.extend([player[0].cost,
ppd * 100000])
mass_hold.append(listify_holder)
check = []
with open('data/fan-pros.csv', 'rb') as csvdata:
for row in csvdata:
check = row
break
with open('data/fan-pros.csv', 'wb') as csvdata:
if len(check) == 4:
pass
else:
writer = csv.writer(csvdata, lineterminator='\n')
writer.writerows(mass_hold)
if league == 'NFL':
_check_missing_players(all_players, args.sp, args.mp)
# filter based on criteria and previously optimized
all_players = filter(lambda x: x.name not in remove and \
x.proj >= int(args.lp) and \
x.cost <= int(args.ms) and \
x.team is not None,
all_players)
variables, solution = run_solver(solver,
all_players,
position_distribution)
if solution == solver.OPTIMAL:
roster = RosterSelect().roster_gen(args.l)
for i, player in enumerate(all_players):
if variables[i].solution_value() == 1:
roster.add_player(player)
print "Optimal roster for: %s" % league
print roster
print
return roster
else:
raise Exception('No solution error')
def run(position_distribution, league, remove, args, test_mode=False):
csv_name = 'test' if test_mode else 'current'
solver = pywraplp.Solver('FD',
pywraplp.Solver.CBC_MIXED_INTEGER_PROGRAMMING)
all_players = []
with open(fns.format(csv_name), 'rb') as csvfile:
csvdata = csv.DictReader(csvfile)
for row in csvdata:
player = Player(row['Position'], row['Name'], row['Salary'],
team=row['teamAbbrev'],
matchup=row['GameInfo'],
lock=(args.locked and
row['Name'] in args.locked))
if args.l == 'NBA':
player.proj = float(row['AvgPointsPerGame'])
player.team = row['teamAbbrev']
all_players.append(player)
if args.w and args.season and args.historical == _YES:
print('Fetching {} season data for all players...'
.format(args.season))
for p in all_players:
p.set_historical(int(args.w), int(args.season))
if league == 'NFL':
if args.po_location and args.po:
with open(args.po_location, 'rb') as csvfile:
csvdata = csv.DictReader(csvfile)
for row in csvdata:
player = filter(
lambda p: p.name in row['Name'],
all_players)
if player:
player[0].projected_ownership_pct = float(row['%'])
with open(fnp.format(csv_name), 'rb') as csvfile:
csvdata = csv.DictReader(csvfile)
mass_hold = [['playername', 'points', 'cost', 'ppd']]
# hack for weird defensive formatting
def name_match(row):
def match_fn(p):
if p.pos == 'DST':
return p.name.strip() in row['playername']
return p.name in row['playername']
return match_fn
for row in csvdata:
player = filter(name_match(row), all_players)
if len(player) == 0:
continue
player[0].proj = float(row['points'])
player[0].marked = 'Y'
listify_holder = [
row['playername'],
row['points']
]
if '0.0' not in row['points'] or player[0].cost != 0:
ppd = float(row['points']) / float(player[0].cost)
else:
ppd = 0
listify_holder.extend([player[0].cost,
ppd * 100000])
mass_hold.append(listify_holder)
check = []
with open(fns.format(csv_name), 'rb') as csvdata:
for row in csvdata:
check = row
break
with open(fnp.format(csv_name), 'wb') as csvdata:
if len(check) == 4:
pass
else:
writer = csv.writer(csvdata, lineterminator='\n')
writer.writerows(mass_hold)
if league == 'NFL':
_check_missing_players(all_players, args.sp, args.mp)
# filter based on criteria and previously optimized
# do not include DST or TE projections in min point threshold.
all_players = filter(
qc.add_constraints(args, remove),
all_players)
if args.no_double_te == _YES:
cons.POSITIONS['NFL'] = cons.get_nfl_positions(te_upper=1)
variables, solution = run_solver(solver,
all_players,
position_distribution,
args)
if solution == solver.OPTIMAL:
roster = RosterSelect().roster_gen(args.l)
for i, player in enumerate(all_players):
if variables[i].solution_value() == 1:
roster.add_player(player)
print "Optimal roster for: %s" % league
print roster
print
return roster
else:
print("No solution found for command line query. " +
"Try adjusting your query by taking away constraints.")
return None
def run(position_distribution, league, remove):
solver = pywraplp.Solver('FD',
pywraplp.Solver.CBC_MIXED_INTEGER_PROGRAMMING)
all_players = []
with open('dk/data/dkk.csv', 'rb') as csvfile:
csvdata = csv.DictReader(csvfile)
for idx, row in enumerate(csvdata):
if idx > 0:
player = Player(row['Position'], row['Name'], row['Salary'],
matchup=row['GameInfo'])
if args.l == 'NBA':
player.proj = float(row['AvgPointsPerGame'])
player.team = row['teamAbbrev']
all_players.append(player)
if league == 'NBA':
with open('dk/data/fan-pros.csv', 'rb') as csvfile:
csvdata = csv.DictReader(csvfile)
mass_hold = [['player', 'projected', 'salary', 'cpp']]
for row in csvdata:
holder = row
player = filter(lambda x: x.name in row['player'], all_players)
if len(player) == 0:
continue
player[0].proj = float(row['projected'])
player[0].marked = 'Y'
player[0].team = row['playername'].split(' ')[-2]
listify_holder = [
row['player'],
row['projected']
]
if '0.0' not in row['projected'] or player[0].cost != 0:
ppd = float(row['projected']) / float(player[0].cost)
else:
ppd = 0
listify_holder.extend([player[0].cost,
ppd * 100000])
mass_hold.append(listify_holder)
check = []
with open('dk/data/fan-pros.csv', 'rb') as csvdata:
for row in csvdata:
check = row
break
with open('dk/data/fan-pros.csv', 'wb') as csvdata:
if len(check) == 4:
pass
else:
writer = csv.writer(csvdata, lineterminator='\n')
writer.writerows(mass_hold)
if league == 'NFL':
_check_missing_players(all_players, args.sp, args.mp)
# filter based on criteria and previously optimized
all_players = filter(lambda x: x.name not in remove and \
x.proj >= int(args.lp) and \
x.cost <= int(args.ms) and \
x.team is not None,
all_players)
variables, solution = run_solver(solver,
all_players,
position_distribution)
if solution == solver.OPTIMAL:
roster = RosterSelect().roster_gen(args.l)
for i, player in enumerate(all_players):
if variables[i].solution_value() == 1:
roster.add_player(player)
print "Optimal roster for: %s" % league
print roster
print
return roster
else:
raise Exception('No solution error')
def run(position_distribution, league, remove, args, test_mode=False):
csv_name = 'test' if test_mode else 'current'
solver = pywraplp.Solver('FD',
pywraplp.Solver.CBC_MIXED_INTEGER_PROGRAMMING)
all_players = []
with open(fns.format(csv_name), 'rb') as csvfile:
csvdata = csv.DictReader(csvfile)
for idx, row in enumerate(csvdata):
if idx > 0:
player = Player(row['Position'], row['Name'], row['Salary'],
matchup=row['GameInfo'])
if args.l == 'NBA':
player.proj = float(row['AvgPointsPerGame'])
player.team = row['teamAbbrev']
all_players.append(player)
if league == 'NFL':
with open(fnp.format(csv_name), 'rb') as csvfile:
csvdata = csv.DictReader(csvfile)
mass_hold = [['playername', 'points', 'cost', 'ppd']]
for row in csvdata:
player = filter(lambda x: x.name in row['playername'],
all_players)
if len(player) == 0:
continue
player[0].proj = float(row['points'])
player[0].marked = 'Y'
player[0].team = row['playername'].split(' ')[-2]
listify_holder = [
row['playername'],
row['points']
]
if '0.0' not in row['points'] or player[0].cost != 0:
ppd = float(row['points']) / float(player[0].cost)
else:
ppd = 0
listify_holder.extend([player[0].cost,
ppd * 100000])
mass_hold.append(listify_holder)
check = []
with open(fns.format(csv_name), 'rb') as csvdata:
for row in csvdata:
check = row
break
with open(fnp.format(csv_name), 'wb') as csvdata:
if len(check) == 4:
pass
else:
writer = csv.writer(csvdata, lineterminator='\n')
writer.writerows(mass_hold)
if league == 'NFL':
_check_missing_players(all_players, args.sp, args.mp)
# filter based on criteria and previously optimized
# do not include DST or TE projections in min point threshold.
if isinstance(args.teams, str):
args.teams = args.teams.split(' ')
all_players = filter(
qc.add_constraints(args, remove),
all_players)
variables, solution = run_solver(solver,
all_players,
position_distribution)
if solution == solver.OPTIMAL:
roster = RosterSelect().roster_gen(args.l)
for i, player in enumerate(all_players):
if variables[i].solution_value() == 1:
roster.add_player(player)
print "Optimal roster for: %s" % league
print roster
print
return roster
else:
print("No solution found for command line query. " +
"Try adjusting your query by taking away constraints.")
return None