Enriched metadata with date, game_week, game_id and coach

kloppy/_providers/statsbomb.py

@@ -18,6 +18,7 @@ def load(
     event_types: Optional[List[str]] = None,
     coordinates: Optional[str] = None,
     event_factory: Optional[EventFactory] = None,
+    additional_metadata: dict = {}
 ) -> EventDataset:
     """
     Load StatsBomb event data into a [`EventDataset`][kloppy.domain.models.event.EventDataset]
@@ -48,6 +49,7 @@ def load(
                 lineup_data=lineup_data_fp,
                 three_sixty_data=three_sixty_data_fp,
             ),
+            additional_metadata=additional_metadata,
         )
 
 

kloppy/domain/models/common.py

@@ -1003,6 +1003,10 @@ class Metadata:
         orientation: See [`Orientation`][kloppy.domain.models.common.Orientation]
         flags:
         provider: See [`Provider`][kloppy.domain.models.common.Provider]
+        date: Date of the game.
+        game_week: Game week (or match day) of the game. It can also be the stage
+        (ex: "8th Finals"), if the game is happening during a cup or a play-off.
+        game_id: Game id of the game from the provider.
     """
 
     teams: List[Team]
@@ -1014,6 +1018,11 @@ class Metadata:
     coordinate_system: CoordinateSystem
     score: Optional[Score] = None
     frame_rate: Optional[float] = None
+    date: Optional[datetime] = None
+    game_week: Optional[str] = None
+    game_id: Optional[str] = None
+    home_coach: Optional[str] = None
+    away_coach: Optional[str] = None
     attributes: Optional[Dict] = field(default_factory=dict, compare=False)
 
     def __post_init__(self):

kloppy/infra/serializers/event/datafactory/deserializer.py

@@ -1,6 +1,7 @@
 import json
 import logging
 from datetime import timedelta, datetime, timezone
+from dateutil.parser import parse, _parser
 from dataclasses import replace
 from typing import Dict, List, Tuple, Union, IO, NamedTuple
 
@@ -453,6 +454,19 @@ def deserialize(self, inputs: DatafactoryInputs) -> EventDataset:
                         periods[half], end_timestamp=timestamp
                     )
 
+            try:
+                date = match["date"]
+                if date:
+                    date = parse(date).astimezone(timezone.utc)
+            except _parser.ParserError:
+                date = None
+            game_week = match.get("week", None)
+            if game_week:
+                game_week = str(game_week)
+            game_id = match.get("matchId", None)
+            if game_id:
+                game_id = str(game_id)
+
             # exclude goals, already listed as shots too
             incidences.pop(DF_EVENT_CLASS_GOALS)
             raw_events = [
@@ -613,6 +627,9 @@ def deserialize(self, inputs: DatafactoryInputs) -> EventDataset:
             score=score,
             provider=Provider.DATAFACTORY,
             coordinate_system=transformer.get_to_coordinate_system(),
+            date=date,
+            game_week=game_week,
+            game_id=game_id,
         )
 
         return EventDataset(

kloppy/infra/serializers/event/sportec/deserializer.py

@@ -83,6 +83,8 @@ class SportecMetadata(NamedTuple):
     x_max: float
     y_max: float
     fps: int
+    home_coach: str
+    away_coach: str
 
 
 def sportec_metadata_from_xml_elm(match_root) -> SportecMetadata:
@@ -99,10 +101,17 @@ def sportec_metadata_from_xml_elm(match_root) -> SportecMetadata:
 
     home_team = away_team = None
     for team_elm in team_elms:
+        head_coach = [
+            trainer.attrib["Shortname"]
+            for trainer in team_elm.TrainerStaff.iterchildren("Trainer")
+            if trainer.attrib["Role"] == "headcoach"
+        ]
         if team_elm.attrib["Role"] == "home":
             home_team = _team_from_xml_elm(team_elm)
+            home_coach = head_coach[0] if len(head_coach) else None
         elif team_elm.attrib["Role"] == "guest":
             away_team = _team_from_xml_elm(team_elm)
+            away_coach = head_coach[0] if len(head_coach) else None
         else:
             raise DeserializationError(
                 f"Unknown side: {team_elm.attrib['Role']}"
@@ -194,6 +203,8 @@ def sportec_metadata_from_xml_elm(match_root) -> SportecMetadata:
         x_max=x_max,
         y_max=y_max,
         fps=SPORTEC_FPS,
+        home_coach=home_coach,
+        away_coach=away_coach,
     )
 
 
@@ -404,12 +415,20 @@ def deserialize(self, inputs: SportecEventDataInputs) -> EventDataset:
             event_root = objectify.fromstring(inputs.event_data.read())
 
         with performance_logging("parse data", logger=logger):
+            date = parse(
+                match_root.MatchInformation.General.attrib["KickoffTime"]
+            ).astimezone(timezone.utc)
+            game_week = match_root.MatchInformation.General.attrib["MatchDay"]
+            game_id = match_root.MatchInformation.General.attrib["MatchId"]
+
             sportec_metadata = sportec_metadata_from_xml_elm(match_root)
             teams = home_team, away_team = sportec_metadata.teams
             transformer = self.get_transformer(
                 pitch_length=sportec_metadata.x_max,
                 pitch_width=sportec_metadata.y_max,
             )
+            home_coach = sportec_metadata.home_coach
+            away_coach = sportec_metadata.away_coach
 
             periods = []
             period_id = 0
@@ -632,6 +651,11 @@ def deserialize(self, inputs: SportecEventDataInputs) -> EventDataset:
             flags=~(DatasetFlag.BALL_STATE | DatasetFlag.BALL_OWNING_TEAM),
             provider=Provider.SPORTEC,
             coordinate_system=transformer.get_to_coordinate_system(),
+            date=date,
+            game_week=game_week,
+            game_id=game_id,
+            home_coach=home_coach,
+            away_coach=away_coach,
         )
 
         return EventDataset(

kloppy/infra/serializers/event/statsbomb/deserializer.py

@@ -36,7 +36,7 @@ class StatsBombDeserializer(EventDataDeserializer[StatsBombInputs]):
     def provider(self) -> Provider:
         return Provider.STATSBOMB
 
-    def deserialize(self, inputs: StatsBombInputs) -> EventDataset:
+    def deserialize(self, inputs: StatsBombInputs, additional_metadata) -> EventDataset:
         # Intialize coordinate system transformer
         self.transformer = self.get_transformer()
 
@@ -118,6 +118,7 @@ def deserialize(self, inputs: StatsBombInputs) -> EventDataset:
             score=None,
             provider=Provider.STATSBOMB,
             coordinate_system=self.transformer.get_to_coordinate_system(),
+            **additional_metadata
         )
         return EventDataset(metadata=metadata, records=events)
 

kloppy/infra/serializers/event/statsperform/deserializer.py

@@ -614,6 +614,9 @@ def deserialize(self, inputs: StatsPerformInputs) -> EventDataset:
             periods = metadata_parser.extract_periods()
             score = metadata_parser.extract_score()
             teams = metadata_parser.extract_lineups()
+            date = events_parser.extract_date()
+            game_week = events_parser.extract_game_week()
+            game_id = events_parser.extract_game_id()
             raw_events = [
                 event
                 for event in events_parser.extract_events()
@@ -827,6 +830,9 @@ def deserialize(self, inputs: StatsPerformInputs) -> EventDataset:
             if inputs.event_feed.upper() == "F24"
             else Provider.STATSPERFORM,
             coordinate_system=transformer.get_to_coordinate_system(),
+            date=date,
+            game_week=game_week,
+            game_id=game_id,
         )
 
         return EventDataset(

kloppy/infra/serializers/event/statsperform/parsers/base.py

@@ -12,7 +12,6 @@
 
 from datetime import datetime
 from dataclasses import dataclass, field
-from typing import List, Optional
 
 
 @dataclass
@@ -53,6 +52,18 @@ def extract_score(self) -> Optional[Score]:
         """Return the score of the game."""
         return None
 
+    def extract_date(self) -> Optional[str]:
+        """Return the date of the game."""
+        return None
+
+    def extract_game_week(self) -> Optional[str]:
+        """Return the game_week of the game."""
+        return None
+
+    def extract_game_id(self) -> Optional[str]:
+        """Return the game_id of the game."""
+        return None
+
     def extract_lineups(self) -> Tuple[Team, Team]:
         """Return the home and away team."""
         raise NotImplementedError

kloppy/infra/serializers/event/statsperform/parsers/f24_xml.py

@@ -1,9 +1,9 @@
 """XML parser for Opta F24 feeds."""
 import pytz
-from datetime import datetime
-from typing import List
+from datetime import datetime, timezone
+from typing import List, Optional
+from dateutil.parser import parse
 
-from kloppy.domain import Period
 from .base import OptaXMLParser, OptaEvent
 
 
@@ -53,3 +53,27 @@ def extract_events(self) -> List[OptaEvent]:
             )
             for event in game_elm.iterchildren("Event")
         ]
+
+    def extract_date(self) -> Optional[str]:
+        """Return the date of the game."""
+        game_elm = self.root.find("Game")
+        if game_elm and "game_date" in game_elm.attrib:
+            return parse(game_elm.attrib["game_date"]).astimezone(timezone.utc)
+        else:
+            return None
+
+    def extract_game_week(self) -> Optional[str]:
+        """Return the game_week of the game."""
+        game_elm = self.root.find("Game")
+        if game_elm and "matchday" in game_elm.attrib:
+            return game_elm.attrib["matchday"]
+        else:
+            return None
+
+    def extract_game_id(self) -> Optional[str]:
+        """Return the game_id of the game."""
+        game_elm = self.root.find("Game")
+        if game_elm and "id" in game_elm.attrib:
+            return game_elm.attrib["id"]
+        else:
+            return None

kloppy/infra/serializers/event/statsperform/parsers/ma1_json.py

@@ -1,6 +1,6 @@
 """JSON parser for Stats Perform MA1 feeds."""
 import pytz
-from datetime import datetime
+from datetime import datetime, timezone
 from typing import Any, Optional, List, Tuple, Dict
 
 from kloppy.domain import Period, Score, Team, Ground, Player
@@ -30,7 +30,13 @@ def extract_periods(self) -> List[Period]:
         return parsed_periods
 
     def extract_score(self) -> Optional[Score]:
-        return None
+        live_data = self.root["liveData"]
+        match_details = live_data["matchDetails"]
+        home_score = match_details["scores"]["total"]["home"]
+        away_score = match_details["scores"]["total"]["away"]
+        if home_score is None or away_score is None:
+            return None
+        return Score(home=home_score, away=away_score)
 
     def extract_lineups(self) -> Tuple[Team, Team]:
         teams = {}
@@ -76,6 +82,29 @@ def extract_lineups(self) -> Tuple[Team, Team]:
             raise DeserializationError("Lineup incomplete")
         return home_team, away_team
 
+    def extract_date(self) -> Optional[str]:
+        """Return the date of the game."""
+        if "matchInfo" in self.root and "date" in self.root["matchInfo"]:
+            return datetime.strptime(
+                self.root["matchInfo"]["date"], "%Y-%m-%dZ"
+            ).astimezone(timezone.utc)
+        else:
+            return None
+
+    def extract_game_week(self) -> Optional[str]:
+        """Return the game_week of the game."""
+        if "matchInfo" in self.root and "week" in self.root["matchInfo"]:
+            return self.root["matchInfo"]["week"]
+        else:
+            return None
+
+    def extract_game_id(self) -> Optional[str]:
+        """Return the game_id of the game."""
+        if "matchInfo" in self.root and "id" in self.root["matchInfo"]:
+            return self.root["matchInfo"]["id"]
+        else:
+            return None
+
     def _parse_teams(self) -> List[Dict[str, Any]]:
         parsed_teams = []
         match_info = self.root["matchInfo"]

kloppy/infra/serializers/event/wyscout/deserializer_v2.py

@@ -499,6 +499,9 @@ def deserialize(self, inputs: WyscoutInputs) -> EventDataset:
                     for wyId, team in teams.items()
                 ]
             )
+            game_id = raw_events["events"][0].get("matchId", None)
+            if game_id:
+                game_id = str(game_id)
 
             events = []
 
@@ -730,6 +733,7 @@ def deserialize(self, inputs: WyscoutInputs) -> EventDataset:
             flags=None,
             provider=Provider.WYSCOUT,
             coordinate_system=transformer.get_to_coordinate_system(),
+            game_id=game_id,
         )
 
         return EventDataset(metadata=metadata, records=events)

kloppy/infra/serializers/event/wyscout/deserializer_v3.py

@@ -1,8 +1,9 @@
 import json
 import logging
 from dataclasses import replace
-from datetime import timedelta
-from typing import Dict, List, Tuple, NamedTuple, IO
+from datetime import timedelta, timezone
+from dateutil.parser import parse
+from typing import Dict, List
 
 from kloppy.domain import (
     BallOutEvent,
@@ -536,6 +537,24 @@ def deserialize(self, inputs: WyscoutInputs) -> EventDataset:
                     for wyId, team in teams.items()
                 ]
             )
+            date = raw_events["match"].get("dateutc")
+            if date:
+                date = parse(date).astimezone(timezone.utc)
+            game_week = raw_events["match"].get("gameweek")
+            if game_week:
+                game_week = str(game_week)
+            game_id = raw_events["events"][0].get("matchId")
+            if game_id:
+                game_id = str(game_id)
+            coaches = raw_events["coaches"]
+            if home_team_id in coaches and "coach" in coaches[home_team_id]:
+                home_coach = coaches[home_team_id]["coach"].get("shortName")
+            else:
+                home_coach = None
+            if away_team_id in coaches and "coach" in coaches[away_team_id]:
+                away_coach = coaches[away_team_id]["coach"].get("shortName")
+            else:
+                away_coach = None
 
             events = []
 
@@ -757,6 +776,11 @@ def deserialize(self, inputs: WyscoutInputs) -> EventDataset:
             flags=None,
             provider=Provider.WYSCOUT,
             coordinate_system=transformer.get_to_coordinate_system(),
+            date=date,
+            game_week=game_week,
+            game_id=game_id,
+            home_coach=home_coach,
+            away_coach=away_coach,
         )
 
         return EventDataset(metadata=metadata, records=events)