def test_load(self):
players = Player.load("6644d767-ab15-4528-a4ce-ae1f8aadb65f", "d97835fd-2e92-4698-8900-1f5abea0a3b6")
assert isinstance(players, dict)
assert len(players) == 2
for key, player in players.items():
assert isinstance(key, str)
assert isinstance(player, Player)
assert key == player.id
def get_similar_player(player,
stat='batting',
num_players=10,
include_shadows=False,
filter_by_postion=True,
filter_dead=True):
"""
Get a list of players that have statlines similar to the selected player.
TODO: Weight the attributes based on importance?
:param player: player to compare
:param stat: stats to compare: 'batting', 'pitching', 'baserunning', 'defense', 'offense'
:param num_players: number of players to return
:param include_shadows: whether to include shadows players in the comparison
:param filter_by_postion: if True, do not include players that are not currently active in this position
:param filter_dead: if True, do not include dead players
:return: dictionary with players as values and relative similarity as the key
"""
bat = [
"thwackability", "divinity", "musclitude", "moxie", "patheticism",
"martyrdom", "tragicness", "buoyancy"
]
pitch = [
"unthwackability", "ruthlessness", "overpowerment", "shakespearianism",
"coldness", "suppression"
]
run = [
"laserlikeness", "continuation", "baseThirst", "indulgence",
"groundFriction"
]
def_ = [
"omniscience", "tenaciousness", "watchfulness", "anticapitalism",
"chasiness"
]
if stat == 'batting':
attrs = bat
position = 'BATTER'
elif stat == 'pitching':
attrs = pitch
position = 'PITCHER'
elif stat == 'baserunning':
attrs = run
position = 'BATTER'
elif stat == 'defense':
attrs = def_
position = 'BATTER'
elif stat == 'offense':
attrs = bat + run
position = 'BATTER'
else:
raise ValueError(f"Invalid stat {stat}")
# Generate table of all players
url = "https://api.sibr.dev/datablase/v1/allPlayers"
if include_shadows:
url += f"?includeShadows={include_shadows}"
all_players = [
reference._apply_type_map(p) for p in requests.get(url).json()
]
if filter_dead:
all_players = [x for x in all_players if x["deceased"] is False]
all_players_table = pandas.DataFrame(
all_players, index=[x["player_id"] for x in all_players])
if filter_by_postion:
all_players_table = all_players_table[
all_players_table["position_type"] == position]
all_table = all_players_table.filter(items=attrs)
min_stat = all_table.min(axis=0)
max_stat = all_table.max(axis=0)
all_scaled = (2.0 * ((all_table - min_stat) / (max_stat - min_stat))) - 1.0
# Generate table row of selected player
player_array = pandas.Series(player.json())
player_array = player_array.filter(items=attrs)
player_scaled = (2.0 * ((player_array - min_stat) /
(max_stat - min_stat))) - 1.0
# PERFORM THE RITUAL
all_norm = all_scaled.apply(np.linalg.norm, axis=1)
player_norm = np.linalg.norm(player_scaled)
results = all_scaled.dot(player_scaled) / (all_norm * player_norm)
# Sort and organize output
results = results.drop(index=player.id, errors='ignore')
sorted_ = results.sort_values(ascending=False)
players = Player.load(*sorted_.index[0:num_players])
players = [players.get(id_) for id_ in sorted_.index[0:num_players]]
return {results[x.id]: x for x in players}
def test_load_bad_id(self):
bad_player = Player.load("6644d767-ab15-4528-a4ce-ae1f8aadb65f", "00000000-0000-0000-0000-000000000000")
assert isinstance(bad_player, dict)
assert len(bad_player) == 1
