Merge pull request #127 from koenvo/some-state-fixes

Minor refactor and some fixes

kloppy/domain/models/common.py

@@ -4,6 +4,7 @@
 from typing import Dict, List, Optional, Callable, Union, Any, TypeVar, Generic
 
 from .pitch import PitchDimensions, Point, Dimension
+from .formation import FormationType
 from ...exceptions import OrientationError
 
 
@@ -144,7 +145,7 @@ class Team:
     team_id: str
     name: str
     ground: Ground
-    starting_formation: str = ""
+    starting_formation: Optional[FormationType] = None
     players: List[Player] = field(default_factory=list)
 
     def __str__(self):

kloppy/domain/models/event.py

@@ -12,6 +12,7 @@
 )
 
 from .common import DataRecord, Dataset, Player, Team
+from .formation import FormationType
 from .pitch import Point
 
 
@@ -130,80 +131,6 @@ class CardType(Enum):
     RED = "RED"
 
 
-class FormationType(Enum):
-    """
-    FormationType
-
-    Attributes:
-        THREE_ONE_TWO_ONE_THREE (str): 3-1-2-1-3 team formation
-        THREE_ONE_THREE_ONE_TWO (str):
-        THREE_ONE_FOUR_TWO (str):
-        THREE_TWO_TWO_ONE_TWO (str):
-        THREE_TWO_TWO_TWO_ONE (str):
-        THREE_TWO_THREE_TWO (str):
-        THREE_THREE_TWO_TWO (str):
-        THREE_TWO_FOUR_ONE (str):
-        THREE_THREE_THREE_ONE (str):
-        THREE_FOUR_ONE_TWO (str):
-        THREE_FOUR_TWO_ONE (str):
-        THREE_FOUR_THREE (str):
-        THREE_FIVE_ONE_ONE (str):
-        THREE_FIVE_TWO (str):
-        FOUR_ONE_TWO_ONE_TWO (str):
-        FOUR_ONE_TWO_TWO_ONE (str):
-        FOUR_ONE_THREE_TWO (str):
-        FOUR_ONE_FOUR_ONE (str):
-        FOUR_TWO_ONE_TWO_ONE (str):
-        FOUR_TWO_TWO_ONE_ONE (str):
-        FOUR_TWO_TWO_TWO (str):
-        FOUR_TWO_THREE_ONE (str):
-        FOUR_TWO_FOUR_ZERO (str):
-        FOUR_THREE_ONE_TWO (str):
-        FOUR_THREE_TWO_ONE (str):
-        FOUR_THREE_THREE (str):
-        FOUR_FOUR_ONE_ONE (str):
-        FOUR_FOUR_TWO (str):
-        FOUR_FIVE_ONE (str):
-        FIVE_TWO_TWO_ONE (str):
-        FIVE_THREE_TWO (str):
-        FIVE_FOUR_ONE (str):
-
-    """
-
-    THREE_ONE_TWO_ONE_THREE = "3-1-2-1-3"
-    THREE_ONE_THREE_ONE_TWO = "3-1-3-1-2"
-    THREE_ONE_FOUR_TWO = "3-1-4-2"
-    THREE_TWO_TWO_ONE_TWO = "3-2-2-1-2"
-    THREE_TWO_TWO_TWO_ONE = "3-2-2-1-2"
-    THREE_TWO_THREE_TWO = "3-2-3-2"
-    THREE_THREE_TWO_TWO = "3-3-2-2"
-    THREE_TWO_FOUR_ONE = "3-2-4-1"
-    THREE_THREE_THREE_ONE = "3-3-3-1"
-    THREE_FOUR_ONE_TWO = "3-4-1-2"
-    THREE_FOUR_TWO_ONE = "3-4-2-1"
-    THREE_FOUR_THREE = "3-4-3"
-    THREE_FIVE_ONE_ONE = "3-5-1-1"
-    THREE_FIVE_TWO = "3-5-2"
-    FOUR_ONE_TWO_ONE_TWO = "4-1-2-1-2"
-    FOUR_ONE_TWO_TWO_ONE = "4-1-2-2-1"
-    FOUR_ONE_THREE_TWO = "4-1-3-2"
-    FOUR_ONE_FOUR_ONE = "4-1-4-1"
-    FOUR_TWO_ONE_TWO_ONE = "4-2-1-2-1"
-    FOUR_TWO_TWO_ONE_ONE = "4-2-2-1-1"
-    FOUR_TWO_TWO_TWO = "4-2-2-2"
-    FOUR_TWO_THREE_ONE = "4-2-3-1"
-    FOUR_TWO_FOUR_ZERO = "4-2-4-0"
-    FOUR_THREE_ONE_TWO = "4-3-1-2"
-    FOUR_THREE_TWO_ONE = "4-3-2-1"
-    FOUR_THREE_THREE = "4-3-3"
-    FOUR_FOUR_ONE_ONE = "4-4-1-1"
-    FOUR_FOUR_TWO = "4-4-2"
-    FOUR_FIVE_ONE = "4-5-1"
-    FIVE_TWO_TWO_ONE = "5-2-2-1"
-    FIVE_THREE_TWO = "5-3-2"
-    FIVE_FOUR_ONE = "5-4-1"
-
-
 class EventType(Enum):
     """
     Attributes:
@@ -653,10 +580,10 @@ class FormationChangeEvent(Event):
     Attributes:
         event_type (EventType): `EventType.FORMATION_CHANGE` (See [`EventType`][kloppy.domain.models.event.EventType])
         event_name (str): `"card"`
-        formation: See [`FormationType`][kloppy.domain.models.event.FormationType]
+        formation_type: See [`FormationType`][kloppy.domain.models.formation.FormationType]
     """
 
-    formation: FormationType
+    formation_type: FormationType
 
     event_type: EventType = EventType.FORMATION_CHANGE
     event_name: str = "formation_change"

kloppy/domain/models/formation.py

@@ -0,0 +1,78 @@
+from enum import Enum
+
+
+class FormationType(Enum):
+    """
+    FormationType
+
+    Attributes:
+        THREE_ONE_TWO_ONE_THREE (str): 3-1-2-1-3 team formation
+        THREE_ONE_THREE_ONE_TWO (str):
+        THREE_ONE_FOUR_TWO (str):
+        THREE_TWO_TWO_ONE_TWO (str):
+        THREE_TWO_TWO_TWO_ONE (str):
+        THREE_TWO_THREE_TWO (str):
+        THREE_THREE_TWO_TWO (str):
+        THREE_TWO_FOUR_ONE (str):
+        THREE_THREE_THREE_ONE (str):
+        THREE_FOUR_ONE_TWO (str):
+        THREE_FOUR_TWO_ONE (str):
+        THREE_FOUR_THREE (str):
+        THREE_FIVE_ONE_ONE (str):
+        THREE_FIVE_TWO (str):
+        FOUR_ONE_TWO_ONE_TWO (str):
+        FOUR_ONE_TWO_TWO_ONE (str):
+        FOUR_ONE_THREE_TWO (str):
+        FOUR_ONE_FOUR_ONE (str):
+        FOUR_TWO_ONE_TWO_ONE (str):
+        FOUR_TWO_TWO_ONE_ONE (str):
+        FOUR_TWO_TWO_TWO (str):
+        FOUR_TWO_THREE_ONE (str):
+        FOUR_TWO_FOUR_ZERO (str):
+        FOUR_THREE_ONE_TWO (str):
+        FOUR_THREE_TWO_ONE (str):
+        FOUR_THREE_THREE (str):
+        FOUR_FOUR_ONE_ONE (str):
+        FOUR_FOUR_TWO (str):
+        FOUR_FIVE_ONE (str):
+        FIVE_TWO_TWO_ONE (str):
+        FIVE_THREE_TWO (str):
+        FIVE_FOUR_ONE (str):
+
+    """
+
+    THREE_ONE_TWO_ONE_THREE = "3-1-2-1-3"
+    THREE_ONE_THREE_ONE_TWO = "3-1-3-1-2"
+    THREE_ONE_FOUR_TWO = "3-1-4-2"
+    THREE_TWO_TWO_ONE_TWO = "3-2-2-1-2"
+    THREE_TWO_TWO_TWO_ONE = "3-2-2-1-2"
+    THREE_TWO_THREE_TWO = "3-2-3-2"
+    THREE_THREE_TWO_TWO = "3-3-2-2"
+    THREE_TWO_FOUR_ONE = "3-2-4-1"
+    THREE_THREE_THREE_ONE = "3-3-3-1"
+    THREE_FOUR_ONE_TWO = "3-4-1-2"
+    THREE_FOUR_TWO_ONE = "3-4-2-1"
+    THREE_FOUR_THREE = "3-4-3"
+    THREE_FIVE_ONE_ONE = "3-5-1-1"
+    THREE_FIVE_TWO = "3-5-2"
+    FOUR_ONE_TWO_ONE_TWO = "4-1-2-1-2"
+    FOUR_ONE_TWO_TWO_ONE = "4-1-2-2-1"
+    FOUR_ONE_THREE_TWO = "4-1-3-2"
+    FOUR_ONE_FOUR_ONE = "4-1-4-1"
+    FOUR_TWO_ONE_TWO_ONE = "4-2-1-2-1"
+    FOUR_TWO_TWO_ONE_ONE = "4-2-2-1-1"
+    FOUR_TWO_TWO_TWO = "4-2-2-2"
+    FOUR_TWO_THREE_ONE = "4-2-3-1"
+    FOUR_TWO_FOUR_ZERO = "4-2-4-0"
+    FOUR_THREE_ONE_TWO = "4-3-1-2"
+    FOUR_THREE_TWO_ONE = "4-3-2-1"
+    FOUR_THREE_THREE = "4-3-3"
+    FOUR_FOUR_ONE_ONE = "4-4-1-1"
+    FOUR_FOUR_TWO = "4-4-2"
+    FOUR_FIVE_ONE = "4-5-1"
+    FIVE_TWO_TWO_ONE = "5-2-2-1"
+    FIVE_THREE_TWO = "5-3-2"
+    FIVE_FOUR_ONE = "5-4-1"
+
+    def __str__(self):
+        return self.value

kloppy/domain/services/state_builder/__init__.py

@@ -3,7 +3,7 @@
 from kloppy.domain import List, EventDataset
 
 # register all of them
-from . import builders
+from . import builders as _builders
 
 from .registered import create_state_builder
 

kloppy/domain/services/state_builder/builders/formation.py

@@ -1,4 +1,5 @@
 from dataclasses import replace, dataclass
+from typing import Optional
 
 from kloppy.domain import (
     FormationChangeEvent,
@@ -12,11 +13,13 @@
 
 @dataclass
 class Formation:
-    home: FormationType
-    away: FormationType
+    home: Optional[FormationType] = None
+    away: Optional[FormationType] = None
 
     def __str__(self):
-        return f"{self.home} {self.away}"
+        if self.home and self.away:
+            return f"{self.home.value} {self.away.value}"
+        return "Unknown"
 
 
 class FormationStateBuilder(StateBuilder):
@@ -33,7 +36,7 @@ def reduce_before(self, state: Formation, event: Event) -> Formation:
     def reduce_after(self, state: Formation, event: Event) -> Formation:
         if isinstance(event, FormationChangeEvent):
             if event.team.ground == Ground.HOME:
-                state = replace(state, home=event.formation.value)
+                state = replace(state, home=event.formation_type)
             else:
-                state = replace(state, away=event.formation.value)
+                state = replace(state, away=event.formation_type)
         return state

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

@@ -12,18 +12,13 @@
     BallState,
     DatasetFlag,
     Orientation,
-    PitchDimensions,
-    Dimension,
     PassEvent,
     ShotEvent,
     TakeOnEvent,
-    CarryEvent,
     GenericEvent,
     PassResult,
     ShotResult,
     TakeOnResult,
-    CarryResult,
-    EventType,
     Ground,
     Score,
     Provider,
@@ -289,7 +284,7 @@ def _parse_formation_change(raw_qualifiers: List) -> Dict:
     formation_id = int(raw_qualifiers[EVENT_QUALIFIER_TEAM_FORMATION])
     formation = formations[formation_id]
 
-    return dict(formation=formation)
+    return dict(formation_type=formation)
 
 
 def _parse_shot(
@@ -349,7 +344,7 @@ def _team_from_xml_elm(team_elm, f7_root) -> Team:
         ground=Ground.HOME
         if team_elm.attrib["Side"] == "Home"
         else Ground.AWAY,
-        starting_formation=formation,
+        starting_formation=FormationType(formation),
     )
     team.players = [
         Player(

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

@@ -355,7 +355,7 @@ def _parse_card(card_dict: Dict) -> Dict:
 def _parse_formation_change(formation_id: int) -> Dict:
     formation = formations[formation_id]
 
-    return dict(formation=formation)
+    return dict(formation_type=formation)
 
 
 def _determine_xy_fidelity_versions(events: List[Dict]) -> Tuple[int, int]:
@@ -413,17 +413,19 @@ def deserialize(self, inputs: StatsbombInputs) -> EventDataset:
                 for player in raw_event["tactics"]["lineup"]
             }
 
-            starting_formations = [
-                str(raw_event["tactics"]["formation"])
+            starting_formations = {
+                raw_event["team"]["id"]: FormationType(
+                    "-".join(list(str(raw_event["tactics"]["formation"])))
+                )
                 for raw_event in raw_events
                 if raw_event["type"]["id"] == SB_EVENT_TYPE_STARTING_XI
-            ]
+            }
 
             home_team = Team(
                 team_id=str(home_lineup["team_id"]),
                 name=home_lineup["team_name"],
                 ground=Ground.HOME,
-                starting_formation="-".join(list(starting_formations[0])),
+                starting_formation=starting_formations[home_lineup["team_id"]],
             )
             home_team.players = [
                 Player(
@@ -440,7 +442,7 @@ def deserialize(self, inputs: StatsbombInputs) -> EventDataset:
                 team_id=str(away_lineup["team_id"]),
                 name=away_lineup["team_name"],
                 ground=Ground.AWAY,
-                starting_formation="-".join(list(starting_formations[1])),
+                starting_formation=starting_formations[away_lineup["team_id"]],
             )
             away_team.players = [
                 Player(
@@ -662,12 +664,13 @@ def deserialize(self, inputs: StatsbombInputs) -> EventDataset:
                     formation_change_event_kwargs = _parse_formation_change(
                         raw_event["tactics"]["formation"]
                     )
-                    event = FormationChangeEvent.create(
+                    formation_change_event = FormationChangeEvent.create(
                         result=None,
                         qualifiers=None,
                         **formation_change_event_kwargs,
                         **generic_event_kwargs,
                     )
+                    new_events.append(formation_change_event)
                 # rest: generic
                 else:
                     generic_event = GenericEvent.create(

kloppy/tests/test_opta.py

@@ -14,6 +14,7 @@
     PassType,
     DatasetType,
     CardType,
+    FormationType,
 )
 
 from kloppy import opta
@@ -52,10 +53,14 @@ def test_correct_deserialization(self, f7_data: str, f24_data: str):
         )
         assert dataset.metadata.teams[0].name == "FC København"
         assert dataset.metadata.teams[0].ground == Ground.HOME
-        assert dataset.metadata.teams[0].starting_formation == "4-4-2"
+        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 == "4-3-3"
+        assert dataset.metadata.teams[1].starting_formation == FormationType(
+            "4-3-3"
+        )
 
         player = dataset.metadata.teams[0].players[0]
         assert player.player_id == "111319"

kloppy/tests/test_state_builder.py

@@ -1,7 +1,7 @@
 import os
 from itertools import groupby
 
-from kloppy.domain import EventType, Event, EventDataset
+from kloppy.domain import EventType, Event, EventDataset, FormationType
 from kloppy.domain.services.state_builder.builder import StateBuilder, T
 from kloppy.utils import performance_logging
 from kloppy import statsbomb
@@ -95,11 +95,15 @@ def test_formation_state_builder(self):
             events_per_formation_change[str(formation)] = len(events)
 
         # inspect FormationChangeEvent usage and formation state_builder
-        assert events_per_formation_change["4-1-4-1"] == 3073
+        assert events_per_formation_change["4-1-4-1"] == 3074
         assert events_per_formation_change["4-4-2"] == 949
 
-        assert dataset.metadata.teams[0].starting_formation == "4-4-2"
-        assert dataset_with_state.events[1].state["formation"].home == "4-4-2"
+        assert dataset.metadata.teams[0].starting_formation == FormationType(
+            "4-4-2"
+        )
+        assert dataset_with_state.events[1].state[
+            "formation"
+        ].home == FormationType("4-4-2")
 
     def test_register_custom_builder(self):
         class CustomStateBuilder(StateBuilder):