def shot_extractor(player_name, match_id, filepath, start_id=0):
"""
Given a player and match name, this function extracts all shots from that match
Args:
- player_name: str, the name of the player as it appears in the data
- match_id: int, the match id which is usually the name of the events json file
Returns:
- shot_dict: dict, a shot dictionary
"""
# define file path
path = filepath + str(match_id) + ".json"
# read events json into a dataframe
df_events = pd.read_json(path)
# initiate an empty dictionary to store shot values
shot_dict = defaultdict(dict)
shot_id = start_id
goal_x1y1 = [120, 36]
goal_x2y2 = [120, 44]
# iterate through the events and extract shot information specific to the player
for i in range(len(df_events)):
try:
if (df_events.player[i]["name"]) == player_name:
if df_events.type[i]["name"] == "Shot":
freeze_frame_list = []
# create a new entry in the dictionary with the key being the id
# add event index to dictionary
shot_dict[shot_id]["index"] = df_events.index[i + 1]
# get the game state at the event index
shot_dict[shot_id]["gamestate"] = game_state_extractor(
df_events.index[i],
match_id=match_id,
player_name=player_name,
filepath=filepath,
)
# reading match_id
shot_dict[shot_id]["match_id"] = match_id
shot_dict[shot_id]["start_X"] = df_events.location[i][0]
shot_dict[shot_id]["start_Y"] = df_events.location[i][1]
shot_dict[shot_id]["end_X"] = df_events.iloc[
i, :]["shot"]["end_location"][0]
shot_dict[shot_id]["end_Y"] = df_events.iloc[
i, :]["shot"]["end_location"][1]
shot_dict[shot_id]["end_Z"] = df_events.iloc[
i, :]["shot"]["end_location"][2]
# shot time
shot_dict[shot_id]["minute"] = df_events.minute[i]
# statsbomb xg information
shot_dict[shot_id]["statsbomb_xg"] = df_events.iloc[
i, :]["shot"]["statsbomb_xg"]
# shot body part
shot_dict[shot_id]["body_part"] = df_events.iloc[
i, :]["shot"]["body_part"]["name"]
# shot outcome
shot_dict[shot_id]["outcome"] = df_events.iloc[
i, :]["shot"]["outcome"]["name"]
# shot type
shot_dict[shot_id]["type"] = df_events.iloc[
i, :]["shot"]["type"]["name"]
for bool_scenario in [
"aerial_won",
"follows_dribble",
"first_time",
"open_goal",
"deflected",
]:
shot_dict[shot_id][
bool_scenario] = shot_exception_handler(
df_events, i, bool_scenario, is_boolean=True)
# *****************************************************************************************
# calculate the pack density for each shot
# ******************************************************************************************
for j in range(
len(df_events.iloc[i, :]["shot"]["freeze_frame"])):
try:
location = df_events.iloc[
i, :]["shot"]["freeze_frame"][j]["location"]
freeze_frame_list.append(
is_inside(
df_events.location[i],
goal_x1y1,
goal_x2y2,
location,
))
except KeyError:
freeze_frame_list.append(1)
shot_dict[shot_id]["pack_density"] = sum(freeze_frame_list)
shot_id += 1
except:
pass
return shot_dict
def pass_extractor(player_name, match_id, filepath, start_id=0):
"""
Given a player and match name, this function extracts all passes from that match
Args:
- player_name: str, the name of the player as it appears in the data
- match_id: int, the match id which is usually the name of the events json file
Returns:
- pass_dict: dict, a pass dictionary
"""
# define file path
path = filepath + str(match_id) + ".json"
# read events json into a dataframe
df_events = pd.read_json(path)
# initiate an empty dictionary to store pass values
pass_dict = defaultdict(dict)
pass_id = start_id
# iterate through the events and extract pass information specific to the player
for i in range(len(df_events)):
try:
if (df_events.player[i]["name"]) == player_name:
if df_events.type[i]["name"] == "Pass":
# create a new entry in the dictionary with the key being the id
# add event index to dictionary
pass_dict[pass_id]["index"] = df_events.index[i + 1]
# get the game state at the event index
pass_dict[pass_id]["gamestate"] = game_state_extractor(
df_events.index[i],
match_id=match_id,
player_name=player_name,
filepath=filepath,
)
# reading match_id
pass_dict[pass_id]["match_id"] = match_id
# then start location
pass_dict[pass_id]["start_X"] = df_events.location[i][0]
pass_dict[pass_id]["start_Y"] = df_events.location[i][1]
# end location
pass_dict[pass_id]["end_X"] = df_events.iloc[i, 17]["end_location"][
0
]
pass_dict[pass_id]["end_Y"] = df_events.iloc[i, 17]["end_location"][
1
]
# pass time
pass_dict[pass_id]["minute"] = df_events.minute[i]
# recipient information
pass_dict[pass_id]["recipient"] = df_events.iloc[i, 17][
"recipient"
]["name"]
# Pass pass length
pass_dict[pass_id]["length"] = df_events.iloc[i, 17]["length"]
# pass angle
pass_dict[pass_id]["angle"] = df_events.iloc[i, 17]["angle"]
# pass height
pass_dict[pass_id]["height_type"] = df_events.iloc[i, 17]["height"][
"name"
]
# *****************************************************************************************
# writing a for loop to check for some scenarios which are boolean (in StatsBomb data, if the
# result is False, it is omitted)
# *****************************************************************************************
for bool_scenario in [
"backheel",
"deflected",
"miscommunication",
"cross",
"cut_back",
"switch",
"shot_assist",
"goal_assist",
]:
pass_dict[pass_id][bool_scenario] = pass_exception_handler(
df_events, i, bool_scenario, is_boolean=True
)
# extract body part information
pass_dict[pass_id]["body_part"] = df_events.iloc[i, 17][
"body_part"
]["name"]
# *****************************************************************************************
# using the exception handler to extract conditional information which is not boolean
# *****************************************************************************************
# first extracting type
pass_dict[pass_id]["type"] = pass_exception_handler(
df_events,
i,
"type",
is_boolean=False,
default_return="Regular Play",
)
# then pass outcome
pass_dict[pass_id]["outcome"] = pass_exception_handler(
df_events,
i,
"outcome",
is_boolean=False,
default_return="Completed",
)
# finally pass technique
pass_dict[pass_id]["technique"] = pass_exception_handler(
df_events,
i,
"technique",
is_boolean=False,
default_return="N/A",
)
# update pass_id for next pass
pass_id += 1
except:
pass
return pass_dict
def intercept_extractor(player_name, match_id, filepath, start_id=0):
"""
Given a player and match name, this function extracts all interceptions from that match
Args:
- player_name: str, the name of the player as it appears in the data
- match_id: int, the match id which is usually the name of the events json file
- filepath: str, location of files
Returns:
- intercept_dict: dict, an interception dictionary
"""
# define file path
path = filepath + str(match_id) + ".json"
# read events json into a dataframe
df_events = pd.read_json(path)
# initiate an empty dictionary to store intercept values
intercept_dict = defaultdict(dict)
intercept_id = start_id
# iterate through the events and extract intercept information specific to the player
for i in range(len(df_events)):
try:
if (df_events.player[i]["name"]) == player_name:
if df_events.type[i]["name"] == "Interception":
# create a new entry in the dictionary with the key being the id
# add event index to dictionary
intercept_dict[intercept_id]["index"] = df_events.index[i +
1]
# get the game state at the event index
intercept_dict[intercept_id][
"gamestate"] = game_state_extractor(
df_events.index[i + 1],
match_id=match_id,
player_name=player_name,
filepath=filepath,
)
# reading match_id
intercept_dict[intercept_id]["match_id"] = match_id
intercept_dict[intercept_id][
"start_X"] = df_events.location[i][0]
intercept_dict[intercept_id][
"start_Y"] = df_events.location[i][1]
intercept_dict[intercept_id]["outcome"] = df_events.iloc[
i, :]["interception"]["outcome"]["name"]
intercept_id += 1
except:
pass
return intercept_dict
def pressure_extractor(player_name, match_id, filepath, start_id=0):
"""
Given a player and match name, this function extracts all pressures from that match
Args:
- player_name: str, the name of the player as it appears in the data
- match_id: int, the match id which is usually the name of the events json file
- filepath: str, location of files
Returns:
- pressure_dict: dict, a pressure dataframe
"""
# define file path
path = filepath + str(match_id) + ".json"
# read events json into a dataframe
df_events = pd.read_json(path)
# initiate an empty dictionary to store pressure values
pressure_dict = defaultdict(dict)
pressure_id = start_id
# iterate through the events and extract pressure information specific to the player
for i in range(len(df_events)):
try:
if (df_events.player[i]["name"]) == player_name:
if df_events.type[i]["name"] == "Pressure":
# create a new entry in the dictionary with the key being the id
# add event index to dictionary
pressure_dict[pressure_id]["index"] = df_events.index[i +
1]
# get the game state at the event index
pressure_dict[pressure_id][
"gamestate"] = game_state_extractor(
df_events.index[i + 1],
match_id=match_id,
player_name=player_name,
filepath=filepath,
)
# reading match_id
pressure_dict[pressure_id]["match_id"] = match_id
pressure_dict[pressure_id]["start_X"] = df_events.location[
i][0]
pressure_dict[pressure_id]["start_Y"] = df_events.location[
i][1]
pressure_dict[pressure_id][
"duration"] = df_events.duration[i]
pressure_dict[pressure_id][
"counterpress"] = pressure_exception_handler(
df_events, i, "counterpress", is_boolean=True)
pressure_id += 1
except:
pass
return pressure_dict
def carry_extractor(player_name, match_id, filepath, start_id=0):
"""
Given a player and match name, this function extracts all carries from that match
Args:
- player_name: str, the name of the player as it appears in the data
- match_id: int, the match id which is usually the name of the events json file
- filepath: str, location of files
Returns:
- carry_dict: dict, a dictionary of carries
"""
# define file path
path = filepath + str(match_id) + ".json"
# read events json into a dataframe
df_events = pd.read_json(path)
# initiate an empty dictionary to store carry values
carry_dict = defaultdict(dict)
carry_id = start_id
# iterate through the events and extract carry information specific to the player
for i in range(len(df_events)):
try:
if (df_events.player[i]["name"]) == player_name:
if df_events.type[i]["name"] == "Carry":
# create a new entry in the dictionary with the key being the id
# add event index to dictionary
carry_dict[carry_id]["index"] = df_events.index[i + 1]
# get the game state at the event index
carry_dict[carry_id]["gamestate"] = game_state_extractor(
df_events.index[i + 1],
match_id=match_id,
player_name=player_name,
filepath=filepath,
)
# reading match_id
carry_dict[carry_id]["match_id"] = match_id
carry_dict[carry_id]["start_X"] = df_events.location[i][0]
carry_dict[carry_id]["start_Y"] = df_events.location[i][1]
carry_dict[carry_id]["end_X"] = df_events.iloc[
i, :]["carry"]["end_location"][0]
carry_dict[carry_id]["end_Y"] = df_events.iloc[
i, :]["carry"]["end_location"][1]
carry_dict[carry_id][
"under_pressure"] = carry_exception_handler(
df_events, i, "under_pressure", is_boolean=True)
carry_dict[carry_id]["angle"] = angle_calculator(
startX=carry_dict[carry_id]["start_X"],
startY=carry_dict[carry_id]["start_Y"],
endX=carry_dict[carry_id]["end_X"],
endY=carry_dict[carry_id]["end_Y"],
)
carry_dict[carry_id]["length"] = length_calculator(
startX=carry_dict[carry_id]["start_X"],
startY=carry_dict[carry_id]["start_Y"],
endX=carry_dict[carry_id]["end_X"],
endY=carry_dict[carry_id]["end_Y"],
)
carry_id += 1
except:
pass
return carry_dict