def test_set_config(self, f24_data: str, f7_data: str):
"""
Make sure the coordinate system can be set and is used
when loading data.
"""
dataset = opta.load(f24_data=f24_data, f7_data=f7_data)
assert isinstance(dataset.metadata.coordinate_system,
KloppyCoordinateSystem)
set_config("coordinate_system", "opta")
dataset = opta.load(f24_data=f24_data, f7_data=f7_data)
assert isinstance(dataset.metadata.coordinate_system,
OptaCoordinateSystem)
def kloppy_load_data(f7, f24):
"""
Args:
f7: filepath to the match details
f24: filepath to the event details
Returns:
events: DataFrame consisting of all events
home_team_id: id of the home team
away_team_id: id of the away team
"""
dataset = opta.load(f7_data=f7, f24_data=f24)
events = dataset.transform(
to_orientation=Orientation.FIXED_HOME_AWAY).to_pandas(
additional_columns={
"event_name":
lambda event: str(getattr(event, "event_name", "")),
"player_name": lambda event: str(getattr(event, "player", "")),
"ball_state":
lambda event: str(getattr(event, "ball_state", "")),
"team_name": lambda event: str(getattr(event, "team", "")),
}, )
metadata = dataset.metadata
home_team, away_team = metadata.teams
return events, home_team.team_id, away_team.team_id
def test_correct_normalized_deserialization(self, f7_data: str,
f24_data: str):
dataset = opta.load(
f24_data=f24_data,
f7_data=f7_data,
)
assert dataset.events[0].coordinates == Point(0.501, 0.506)
def test_to_pandas_generic_events(self):
base_dir = os.path.dirname(__file__)
dataset = opta.load(
f7_data=f"{base_dir}/files/opta_f7.xml",
f24_data=f"{base_dir}/files/opta_f24.xml",
)
dataframe = dataset.to_pandas()
dataframe = dataframe[dataframe.event_type == "BALL_OUT"]
assert dataframe.shape[0] == 2
def test_config_context(self, f24_data: str, f7_data: str):
assert get_config("coordinate_system") == "kloppy"
with config_context("coordinate_system", "opta"):
assert get_config("coordinate_system") == "opta"
dataset = opta.load(f24_data=f24_data, f7_data=f7_data)
assert get_config("coordinate_system") == "kloppy"
assert isinstance(dataset.metadata.coordinate_system,
OptaCoordinateSystem)
def test_custom_adapter(self, f24_data: str, f7_data: str):
"""
Make sure the coordinate system can be set and is used
when loading data.
"""
class CustomAdapter(Adapter):
def supports(self, url: str) -> bool:
return url.startswith("test123://")
def read_to_stream(self, url: str, output: BinaryIO):
if url == "test123://f24":
fp = open(f24_data, "rb")
elif url == "test123://f7":
fp = open(f7_data, "rb")
else:
raise Exception(f"Unknown url {url}")
output.write(fp.read())
fp.close()
with config_context("cache", None):
with pytest.raises(InputNotFoundError):
# When our adapter is not registered yet
# kloppy will fall back to regular `open`.
opta.load(f24_data="test123://f24", f7_data="test123://f7")
custom_adapter = CustomAdapter()
adapters.append(custom_adapter)
dataset = opta.load(f24_data="test123://f24",
f7_data="test123://f7")
# Asserts borrowed from `test_opta.py`
assert dataset.metadata.provider == Provider.OPTA
assert dataset.dataset_type == DatasetType.EVENT
assert len(dataset.events) == 20
# cleanup
adapters.remove(custom_adapter)
def run_query(argv=sys.argv[1:]):
parser = argparse.ArgumentParser(description="Run query on event data")
parser.add_argument(
"--input-statsbomb",
help="StatsBomb event input files (events.json,lineup.json)",
)
parser.add_argument("--input-opta",
help="Opta event input files (f24.xml,f7.xml)")
parser.add_argument("--input-datafactory",
help="Datafactory event input file (.json)")
parser.add_argument("--input-wyscout", help="Wyscout event input file")
parser.add_argument("--output-xml", help="Output file")
parser.add_argument(
"--with-success",
default=True,
help="Input existence of success capture in output",
)
parser.add_argument("--prepend-time",
default=7,
help="Seconds to prepend to match")
parser.add_argument("--append-time",
default=5,
help="Seconds to append to match")
parser.add_argument("--query-file",
help="File containing the query",
required=True)
parser.add_argument(
"--stats",
default="none",
help="Show matches stats",
choices=["text", "json", "none"],
)
parser.add_argument(
"--show-events",
default=False,
help="Show events for each match",
action="store_true",
)
parser.add_argument(
"--only-success",
default=False,
help="Only show/output success cases",
action="store_true",
)
logger = logging.getLogger("run_query")
logging.basicConfig(
stream=sys.stderr,
level=logging.INFO,
format="%(asctime)s - %(name)s - %(levelname)s - %(message)s",
)
opts = parser.parse_args(argv)
query = load_query(opts.query_file)
dataset = None
if opts.input_statsbomb:
with performance_logging("load dataset", logger=logger):
events_filename, lineup_filename = opts.input_statsbomb.split(",")
dataset = statsbomb.load(
event_data=events_filename.strip(),
lineup_data=lineup_filename.strip(),
event_types=query.event_types,
)
if opts.input_opta:
with performance_logging("load dataset", logger=logger):
f24_filename, f7_filename = opts.input_opta.split(",")
dataset = opta.load(
f24_data=f24_filename.strip(),
f7_data=f7_filename.strip(),
event_types=query.event_types,
)
if opts.input_datafactory:
with performance_logging("load dataset", logger=logger):
events_filename = opts.input_datafactory
dataset = datafactory.load(
event_data=events_filename.strip(),
event_types=query.event_types,
)
if opts.input_wyscout:
with performance_logging("load dataset", logger=logger):
events_filename = opts.input_wyscout
dataset = wyscout.load(
event_data=events_filename,
event_types=query.event_types,
)
if not dataset:
raise Exception("You have to specify a dataset.")
with performance_logging("searching", logger=logger):
matches = pm.search(dataset, query.pattern)
# Construct new code dataset with same properties (eg periods)
# as original event dataset.
# Records will be added later below
records = []
counter = Counter()
for i, match in enumerate(matches):
team = match.events[0].team
success = "success" in match.captures
counter.update({
f"{team.ground}_total": 1,
f"{team.ground}_success": 1 if success else 0,
})
should_process = not opts.only_success or success
if opts.show_events and should_process:
print_match(i, match, success, str(team))
if opts.output_xml and should_process:
code_ = str(team)
if opts.with_success and success:
code_ += " success"
code = Code(
period=match.events[0].period,
code_id=str(i),
code=code_,
timestamp=match.events[0].timestamp - opts.prepend_time,
end_timestamp=match.events[-1].timestamp + opts.append_time,
# TODO: refactor those two out
ball_state=None,
ball_owning_team=None,
)
records.append(code)
code_dataset = CodeDataset(metadata=dataset.metadata, records=records)
if opts.output_xml:
sportscode.save(code_dataset, opts.output_xml)
logger.info(f"Wrote {len(code_dataset.codes)} video fragments to file")
if opts.stats == "text":
text_stats = """\
Home:
total count: {home_total}
success: {home_success} ({home_success_rate:.0f}%)
no success: {home_failure} ({home_failure_rate:.0f}%)
Away:
total count: {away_total}
success: {away_success} ({away_success_rate:.0f}%)
no success: {away_failure} ({away_failure_rate:.0f}%)
""".format(
home_total=counter["home_total"],
home_success=counter["home_success"],
home_success_rate=(counter["home_success"] /
counter["home_total"] * 100),
home_failure=counter["home_total"] - counter["home_success"],
home_failure_rate=(
(counter["home_total"] - counter["home_success"]) /
counter["home_total"] * 100),
away_total=counter["away_total"],
away_success=counter["away_success"],
away_success_rate=(counter["away_success"] /
counter["away_total"] * 100),
away_failure=counter["away_total"] - counter["away_success"],
away_failure_rate=(
(counter["away_total"] - counter["away_success"]) /
counter["away_total"] * 100),
)
print(textwrap.dedent(text_stats))
elif opts.stats == "json":
import json
print(json.dumps(counter, indent=4))
def test_correct_deserialization(self, f7_data: str, f24_data: str):
dataset = opta.load(f24_data=f24_data,
f7_data=f7_data,
coordinates="opta")
assert dataset.metadata.provider == Provider.OPTA
assert dataset.dataset_type == DatasetType.EVENT
assert len(dataset.events) == 20
assert len(dataset.metadata.periods) == 2
assert (
dataset.events[10].ball_owning_team == dataset.metadata.teams[1]
) # 1594254267
assert (
dataset.events[15].ball_owning_team == dataset.metadata.teams[0]
) # 2087733359
assert (
dataset.metadata.orientation == Orientation.ACTION_EXECUTING_TEAM)
assert dataset.metadata.teams[0].name == "FC København"
assert dataset.metadata.teams[0].ground == Ground.HOME
assert dataset.metadata.teams[0].starting_formation == FormationType(
"4-4-2")
assert dataset.metadata.teams[1].name == "FC Nordsjælland"
assert dataset.metadata.teams[1].ground == Ground.AWAY
assert dataset.metadata.teams[1].starting_formation == FormationType(
"4-3-3")
player = dataset.metadata.teams[0].players[0]
assert player.player_id == "111319"
assert player.jersey_no == 21
assert str(player) == "Jesse Joronen"
assert player.position.position_id == "1"
assert player.position.name == "Goalkeeper"
assert dataset.metadata.periods[0] == Period(
id=1,
start_timestamp=1537714933.608,
end_timestamp=1537717701.222,
attacking_direction=AttackingDirection.NOT_SET,
)
assert dataset.metadata.periods[1] == Period(
id=2,
start_timestamp=1537718728.873,
end_timestamp=1537721737.788,
attacking_direction=AttackingDirection.NOT_SET,
)
assert dataset.events[0].coordinates == Point(50.1, 49.4)
# Check the qualifiers
assert (dataset.events[0].qualifiers[0].value == SetPieceType.KICK_OFF
) # 1510681159
assert (dataset.events[6].qualifiers[0].value == BodyPart.HEAD
) # 1101592119
assert (dataset.events[5].qualifiers[0].value == PassType.CHIPPED_PASS
) # 1444075194
assert (dataset.events[19].qualifiers[0].value == CardType.RED
) # 2318695229
# Check receiver coordinates for incomplete passes
assert dataset.events[6].receiver_coordinates.x == 45.5
assert dataset.events[6].receiver_coordinates.y == 68.2
# Check timestamp from qualifier in case of goal
assert dataset.events[17].timestamp == 139.65200018882751 # 2318695229
# assert dataset.events[17].coordinates_y == 12
# Check Own goal
assert dataset.events[18].result.value == "OWN_GOAL" # 2318697001
