def test_deterministic_exposure_limits():
iterations = 2
players = salary_download.generate_players_from_csvs(
salary_file_location=salary_file,
projection_file_location=projection_file,
game=rules.DRAFT_KINGS,
)
rosters, exposure_diffs = run_multi(iterations=2,
rule_set=rules.DK_NFL_RULE_SET,
player_pool=players,
exposure_bounds=[
{
'name': 'Andrew Luck',
'min': 0.5,
'max': 0.7
},
{
'name': 'Alshon Jeffery',
'min': 1,
'max': 1
},
])
ntools.assert_equal(len(rosters), iterations)
ntools.assert_equal(len(exposure_diffs), 0)
players = [p.name for p in rosters[0].players]
ntools.assert_true('Andrew Luck' in players)
ntools.assert_true('Alshon Jeffery' in players)
players = [p.name for p in rosters[1].players]
ntools.assert_true('Andrew Luck' not in players)
ntools.assert_true('Alshon Jeffery' in players)
def test_random_exposure_limits():
iterations = 10
players = salary_download.generate_players_from_csvs(
salary_file_location=salary_file,
projection_file_location=projection_file,
game=rules.DRAFT_KINGS,
)
rosters, exposure_diffs = run_multi(
iterations=iterations,
exposure_random_seed=42,
rule_set=rules.DK_NFL_RULE_SET,
player_pool=players,
)
ntools.assert_equal(len(rosters), iterations)
ntools.assert_equal(len(exposure_diffs), 0)
def Optimize(date, top_k, slate_id, projection_source = 'rotogrinders'):
projections_path = GetMetaDataPath() + '/rotogrinders/projections/%s.csv' % (date)
if not os.path.exists(projections_path):
print("projections for %s not exists" % date)
return
projections = pd.read_csv(projections_path, header = 0, index_col = 0)
slates_path = GetMetaDataPath() + '/draftkings/slates/%s.csv' % (date)
if not os.path.exists(slates_path):
print("slates for %s not exists" % date)
return
slates = pd.read_csv(slates_path, header = 0, index_col = 0)
if slate_id not in slates['SLATE_ID'].tolist():
print("slate %d not exist!" % slate_id)
return
if slates[slates['SLATE_ID'] == slate_id]['SLATE_TYPE'].values[0] != 'CLASSIC':
print("slate type %s not supported!" % slate_type)
return
players_path = GetMetaDataPath() + '/draftkings/players/%s.csv' % (date)
players = pd.read_csv(players_path, header = 0, index_col = 0)
players = players[players['SLATE_ID'] == slate_id]
players.set_index('NAME', inplace=True)
players_with_projection = projections.join(players, how='inner')
player_pool = []
for name, data in players_with_projection.iterrows():
for pos in data['POSITION'].split('/'):
player_pool.append(Player(
name=name,
cost=data['SALARY'],
proj=data['PROJECTION'],
pos=pos,
possible_positions=data['POSITION'],
multi_position='/' in data['POSITION'],
team=data['TEAM'],
matchup=data['GAME_INFO'],
average_score=data['DKP'],))
roster = run_multi(
iterations=top_k,
rule_set=rules.DK_NBA_RULE_SET,
player_pool=player_pool,
verbose=True,
)
def _generate_rosters(self) -> Roster:
players = self._generate_players()
custom_rules = self._set_additional_rules()
try:
constraints, exposures, optimizers = self._compile_params()
except NameError as e:
raise ("Constrains error {}".format(e))
try:
return run_multi(
iterations=self.entries,
exposure_bounds=exposures,
rule_set=custom_rules,
player_pool=players,
optimizer_settings=optimizers,
constraints=constraints,
verbose=True,
)
except InvalidBoundsException as e:
raise ("Error: {}".format(e))
def optimizeJSON(payload):
raw = pd.read_json((payload["raw"]))
data = pd.read_json((payload["posted"]))
optimizer = None
if (payload["site"] == "DRAFTKINGS"):
if (payload["sport_type"] == "PGA"):
optimizer = get_optimizer(Site.DRAFTKINGS, Sport.GOLF)
elif (payload["sport_type"] == "NFL"):
optimizer = get_optimizer(Site.DRAFTKINGS, Sport.FOOTBALL)
else:
optimizer = get_optimizer(Site.DRAFTKINGS, Sport.BASKETBALL)
raw['AvgPointsPerGame'] = raw[payload["projection"]]
raw['AvgPointsPerGame'] = pd.to_numeric(raw["AvgPointsPerGame"]).fillna(0)
else:
if (payload["sport_type"] != "PGA"):
optimizer = get_optimizer(Site.FANDUEL, Sport.FOOTBALL)
if (payload["sport_type"] == "NBA"):
optimizer = get_optimizer(Site.FANDUEL, Sport.BASKETBALL)
raw['FPPG'] = raw[payload["projection"]]
raw['FPPG'] = pd.to_numeric(raw["FPPG"]).fillna(0)
raw['First Name'] = raw['Name']
raw['Last Name'] = raw['Name']
raw['Nickname'] = raw['Name']
raw["Injury Indicator"] = ""
raw['Team'] = raw['TeamAbbrev']
# if os.path.exists("lineup.csv"):
# os.remove("lineup.csv")
lineup_writer = open('/home/johnh/lineup.csv', 'w+')
lineup_writer.close()
raw.to_csv("/home/johnh/lineup.csv")
del raw
if payload["site"] == "FANDUEL" and payload["sport_type"] == "PGA":
players = salary_download.generate_players_from_csvs(
salary_file_location='/home/johnh/lineup.csv',
game=rules.FAN_DUEL,
)
iterations=payload["lineups"]
index = 0
exposure_players = []
locked = []
removed = []
for r in data["Name"]:
if (data["Lock"][index] == 0) and (data["Remove"][index] == 0):
exp_player = {}
exp_player["name"] = r
exp_player["min"] = 0.1
exp_player["max"] = data["Exposure"][index] / 100.0
exposure_players.append(exp_player)
if (data["Lock"][index] == 1) and (data["Remove"][index] == 0):
locked.append(data["Name"][index])
if (data["Lock"][index] == 0) and (data["Remove"][index] == 1):
removed.append(data["Name"][index])
index = index + 1
rosters, exposure_diffs = run_multi(
iterations=iterations,
rule_set=rules.FD_PGA_RULE_SET,
player_pool=players,
exposure_bounds=exposure_players,
constraints=LineupConstraints(
locked=locked,
banned=removed
)
)
ntools.assert_equal(len(rosters), iterations)
ntools.assert_equal(len(exposure_diffs), 0)
lineups = []
exports = []
exports.append(["G","G","G","G","G","G","Budget","FPPG"])
for r in rosters:
export = []
for p in r.sorted_players():
lineup = []
lineup.append("G")
lineup.append(p.name)
lineup.append(p.team)
lineup.append(p.proj)
lineup.append(p.cost)
lineups.append(lineup)
name_index = list(data["Name"]).index(p.name)
export.append(data["ID"][name_index] + ":" + p.name)
total = ["", "", ""]
total.append("{0:.2f}".format(round(r.projected(),2)))
total.append(r.spent())
lineups.append(total)
exports.append(export)
del lineups
del exports
del total
del iterations
del exposure_players
del locked
del removed
result = {"lineups": lineups,
"export": exports}
return result
optimizer.load_players_from_csv("/home/johnh/lineup.csv")
# advanced options (lock, remove and exposure)
for index, row in data.iterrows():
if (row['Remove'] == 1):
print("Removing player: " + row['Name'], row['Remove'])
player = optimizer.get_player_by_id(row['ID'])
optimizer.remove_player(player)
if (row['Lock'] == 1):
print("Lock player: " + row['Name'], row['Remove'])
player = optimizer.get_player_by_id(row['ID'])
optimizer.add_player_to_lineup(player)
if ((row['Exposure'] >= 0) and (row['Exposure'] < 100)):
#print("Set exposure : " + row['Name'], row['Exposure'] / 100.0)
player = optimizer.get_player_by_id(row['ID'])
#print(player)
player.max_exposure = row['Exposure'] / 100.0
if (payload["sport_type"] == "NFL") and (payload["site"] == "DRAFTKINGS"):
optimizer.set_min_salary_cap(49500)
lineupGenerator = optimizer.optimize(payload["lineups"])
def render_player(player):
# type: (LineupPlayer) -> str
result = player.full_name
#print player.id
if player.id:
# if payload["site"] == "FANDUEL" and payload["sport_type"] == "NFL":
# result = player.id + ":" + result
# else:
result += '(%s)' % player.id
return result
rows = []
header = None
print(lineupGenerator)
for index, lineup in enumerate(lineupGenerator):
if index == 0:
header = [player.lineup_position for player in lineup.lineup]
header.extend(('Budget', 'FPPG'))
row = [(render_player)(player) for player in lineup.lineup]
row.append(str(lineup.salary_costs))
row.append(str(lineup.fantasy_points_projection))
rows.append(row)
data = pd.DataFrame.from_records(rows, columns = header)
lineups = pd.DataFrame()
names = []
positions = []
teams = []
fppgs = []
salaries = []
# add number to duplicate cols
def rename_duplicates(old):
new = []
seen = {}
for x in old:
if x in seen:
seen[x] += 1
new.append( "%s_%d" % (x, seen[x]))
else:
seen[x] = 0
new.append(x)
return new
data.columns = rename_duplicates(list(data.columns))
for index, row in data.iterrows():
for player in data.columns[:-2]:
playerName = row[player].split("(")[0]
names.append(playerName)
positions.append(player)
player = optimizer.get_player_by_name(playerName)
fppgs.append(player.fppg)
teams.append(player.team)
salaries.append(player.salary)
positions.append("Totals")
names.append("")
teams.append("")
fppgs.append(row["FPPG"])
salaries.append(row["Budget"])
positions.append("")
names.append("")
teams.append("")
fppgs.append("")
salaries.append("")
# print row["PG"]
lineups["Positions"] = positions
lineups["Name"] = names
lineups["Team"] = teams
lineups["Projection"] = fppgs
lineups["salaries"] = salaries
result = {"lineups": lineups.values.tolist(),
"export": [list(data.columns)] + data.values.tolist()}
del payload
return jsonify(result)
from draftfast.lineup_contraints import LineupConstraints
import csv
player_pool = []
with open('input.csv') as inputfile:
csvreader = csv.reader(inputfile)
isHeader = True
for row in csvreader:
if isHeader:
minSalary = float(row[0])
maxSalary = float(row[1])
lineups = int(row[2])
isHeader = False
else:
player = Player(name=row[1] + '(' + row[0] + ')', cost=float(row[2]), proj=float(row[5]), pos='D')
player_pool.append(player)
player_settings = PlayerPoolSettings(min_salary=minSalary, max_salary=maxSalary)
rosters = run_multi(lineups, rule_set=rules.FD_NASCAR_RULE_SET, player_pool=player_pool, verbose=True)
with open('output.csv', 'w') as outputfile:
csvwriter = csv.writer(outputfile)
for roster in rosters[0]:
players = roster.players
cost = players[0].cost + players[1].cost + players[2].cost + players[3].cost + players[4].cost + players[5].cost
csvwriter.writerow([players[0].name, players[1].name, players[2].name, players[3].name, players[4].name, players[5].name, cost])
the filenames listed in the script.
"""
downloads = environ.get('downloads')
players = salary_download.generate_players_from_csvs(
salary_file_location='{}/NBA_SALS.csv'.format(downloads),
projection_file_location='{}/NBA_PROJECTIONS.csv'.format(downloads),
game=rules.DRAFT_KINGS,
)
rosters, _ = run_multi(
iterations=20,
exposure_bounds=[
{
'name': 'Damian Lillard',
'min': 0.3,
'max': 0.5,
},
{
'name': 'LaMarcus Aldridge',
'min': 0.5,
'max': 0.7,
},
],
rule_set=rules.DK_NBA_RULE_SET,
player_pool=players,
verbose=True,
)
uploader = uploaders.DraftKingsNBAUploader(
pid_file='{}/NBA_PIDS.csv'.format(downloads), )
uploader.write_rosters(rosters)