[Wyscout v3] Estimate shot result coordinates

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

@@ -2,7 +2,7 @@
 import logging
 from dataclasses import replace
 from datetime import timedelta
-from typing import Dict, List, Tuple, NamedTuple, IO
+from typing import Dict, List, Tuple, NamedTuple, IO, Optional
 
 from kloppy.domain import (
     BallOutEvent,
@@ -49,6 +49,7 @@
 
 from ..deserializer import EventDataDeserializer
 from .deserializer_v2 import WyscoutInputs
+from . import wyscout_tags
 
 
 logger = logging.getLogger(__name__)
@@ -99,6 +100,82 @@ def _parse_team(raw_events, wyId: str, ground: Ground) -> Team:
     return team
 
 
+def _create_shot_result_coordinates(raw_event: Dict) -> Optional[Point]:
+    """Estimate the shot end location from the Wyscout tags.
+
+    Wyscout does not provide end-coordinates of shots. Instead shots on goal
+    are tagged with a zone. This function maps each of these zones to
+    a coordinate. The zones and corresponding y-coordinate are depicted below.
+
+
+        olt      | ot |      ort
+     --------------------------------
+          ||=================||
+     -------------------------------
+          || g;l | gt | grt ||
+     --------------------------------
+      ol || gcl | gc  | gcr || or
+     --------------------------------
+      olb || glb  | gb | gln || grb
+
+      40     45    50    55     60    (y-coordinate of zone)
+        44.62               55.38     (y-coordiante of post)
+    """
+    if (
+        raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.GoalBottomCenter
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.GoalCenter
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.GoalTopCenter
+    ):
+        return Point(100.0, 50.0)
+    if (
+        raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.GoalBottomRight
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.GoalCenterRight
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.GoalTopRight
+    ):
+        return Point(100.0, 55.0)
+    if (
+        raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.GoalBottomLeft
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.GoalCenterLeft
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.GoalTopLeft
+    ):
+        return Point(100.0, 45.0)
+    if raw_event["shot"]["goalZone"] == wyscout_tags.ShotZoneResults.OutTop:
+        return Point(100.0, 50.0)
+    if (
+        raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.OutRightTop
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.OutRight
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.OutBottomRight
+    ):
+        return Point(100.0, 60.0)
+    if (
+        raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.OutLeftTop
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.OutLeft
+        or raw_event["shot"]["goalZone"]
+        == wyscout_tags.ShotZoneResults.OutBottomLeft
+    ):
+        return Point(100.0, 40.0)
+    if raw_event["shot"]["goalZone"] == wyscout_tags.ShotZoneResults.Blocked:
+        return Point(
+            x=float(raw_event["location"]["x"]),
+            y=float(raw_event["positions"]["y"]),
+        )
+    return None
+
+
 def _generic_qualifiers(raw_event: Dict) -> List[Qualifier]:
     qualifiers: List[Qualifier] = []
 
@@ -131,10 +208,7 @@ def _parse_shot(raw_event: Dict) -> Dict:
 
     return {
         "result": result,
-        "result_coordinates": Point(
-            x=float(0),
-            y=float(0),
-        ),
+        "result_coordinates": _create_shot_result_coordinates(raw_event),
         "qualifiers": qualifiers,
     }
 
@@ -187,12 +261,14 @@ def _parse_pass(raw_event: Dict, next_event: Dict, team: Team) -> Dict:
         "qualifiers": _pass_qualifiers(raw_event),
         "receive_timestamp": None,
         "receiver_player": receiver_player,
-        "receiver_coordinates": Point(
-            x=float(raw_event["pass"]["endLocation"]["x"]),
-            y=float(raw_event["pass"]["endLocation"]["y"]),
-        )
-        if len(raw_event["pass"]["endLocation"]) > 1
-        else None,
+        "receiver_coordinates": (
+            Point(
+                x=float(raw_event["pass"]["endLocation"]["x"]),
+                y=float(raw_event["pass"]["endLocation"]["y"]),
+            )
+            if len(raw_event["pass"]["endLocation"]) > 1
+            else None
+        ),
     }
 
 
@@ -516,9 +592,11 @@ def deserialize(self, inputs: WyscoutInputs) -> EventDataset:
                     periods.append(
                         Period(
                             id=period_id,
-                            start_timestamp=timedelta(seconds=0)
-                            if len(periods) == 0
-                            else periods[-1].end_timestamp,
+                            start_timestamp=(
+                                timedelta(seconds=0)
+                                if len(periods) == 0
+                                else periods[-1].end_timestamp
+                            ),
                             end_timestamp=None,
                         )
                     )
@@ -539,20 +617,26 @@ def deserialize(self, inputs: WyscoutInputs) -> EventDataset:
                     ball_owning_team = teams[
                         str(raw_event["possession"]["team"]["id"])
                     ]
+                else:
+                    ball_owning_team = team  # TODO: this solve the issue of ball owning team when transforming to spald, but it's not correct
 
                 generic_event_args = {
                     "event_id": raw_event["id"],
                     "raw_event": raw_event,
-                    "coordinates": Point(
-                        x=float(raw_event["location"]["x"]),
-                        y=float(raw_event["location"]["y"]),
-                    )
-                    if raw_event["location"]
-                    else None,
+                    "coordinates": (
+                        Point(
+                            x=float(raw_event["location"]["x"]),
+                            y=float(raw_event["location"]["y"]),
+                        )
+                        if raw_event["location"]
+                        else None
+                    ),
                     "team": team,
-                    "player": players[team_id][player_id]
-                    if player_id != INVALID_PLAYER
-                    else None,
+                    "player": (
+                        players[team_id][player_id]
+                        if player_id != INVALID_PLAYER
+                        else None
+                    ),
                     "ball_owning_team": ball_owning_team,
                     "ball_state": None,
                     "period": periods[-1],

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

@@ -1,3 +1,5 @@
+from enum import Enum
+
 GOAL = 101
 OWN_GOAL = 102
 
@@ -87,3 +89,23 @@
 DANGEROUS_BALL_LOST = 2001
 
 BLOCKED = 2101
+
+
+class ShotZoneResults(Enum):
+    GoalBottomLeft = "glb"
+    GoalBottomRight = "gbr"
+    GoalBottomCenter = "gbc"
+    GoalCenterLeft = "gcl"
+    GoalCenter = "gc"
+    GoalCenterRight = "gcr"
+    GoalTopLeft = "gtl"
+    GoalTopRight = "gtr"
+    GoalTopCenter = "gtc"
+    OutBottomRight = "obr"
+    OutBottomLeft = "obl"
+    OutRight = "or"
+    OutLeft = "ol"
+    OutLeftTop = "olt"
+    OutTop = "ot"
+    OutRightTop = "ort"
+    Blocked = "bc"