Added to all event data providers (#115)

databallpy/load_data/event_data/_normalize_playing_direction_events.py

@@ -4,7 +4,7 @@
 def _normalize_playing_direction_events(
     event_data: pd.DataFrame, home_team_id: int, away_team_id: int
 ) -> pd.DataFrame:
-    """Function to normalize the playing direction based on shots, goals and misses so
+    """Function to normalize the playing direction based on databallpy shots so
     that the home team is always represented as playing from left to right and the
     away team from right to left.
 
@@ -16,14 +16,15 @@ def _normalize_playing_direction_events(
     Returns:
         pd.DataFrame: normalized event data.
     """
-    events = ["shot", "goal", "miss"]
+
     to_changes_cols = [xy for xy in event_data.columns if "_x" in xy or "_y" in xy]
     for period in event_data["period_id"].unique():
         temp_ed = event_data[event_data["period_id"] == period]
 
         # home team
         home = temp_ed.loc[
-            (temp_ed["team_id"] == home_team_id) & (temp_ed["event"].isin(events))
+            (temp_ed["team_id"] == home_team_id)
+            & (temp_ed["databallpy_event"] == "shot")
         ]
         if home["start_x"].mean() < 0:  # home players shoot on the left goal
             event_data.loc[
@@ -32,7 +33,8 @@ def _normalize_playing_direction_events(
 
         # away team
         away = temp_ed.loc[
-            (temp_ed["team_id"] == away_team_id) & (temp_ed["event"].isin(events))
+            (temp_ed["team_id"] == away_team_id)
+            & (temp_ed["databallpy_event"] == "shot")
         ]
         if away["start_x"].mean() > 0:  # away players shoot on the right goal
             event_data.loc[

databallpy/load_data/event_data/instat.py

@@ -9,6 +9,30 @@
 from databallpy.utils.tz_modification import utc_to_local_datetime
 from databallpy.utils.utils import MISSING_INT
 
+instat_databallpy_map = {
+    "Attacking pass accurate": "pass",
+    "Attacking pass inaccurate": "pass",
+    "Unsuccessful dribbling": "dribble",
+    "Successful dribbling": "dribble",
+    "Dribbling": "dribble",
+    "Inaccurate key pass": "pass",
+    "Crosses inaccurate": "pass",
+    "Blocked shot": "shot",
+    "Shots blocked": "shot",
+    "Wide shot": "shot",
+    "Accurate crossing from set piece with a shot": "pass",
+    "Shot on target": "shot",
+    "Crosses accurate": "pass",
+    "Pass into offside": "pass",
+    "Accurate key pass": "pass",
+    "Shot blocked by field player": "shot",
+    "Inaccurate set-piece cross": "pass",
+    "Accurate crossing from set piece": "pass",
+    "Key assist": "pass",
+    "Goal": "shot",
+    "Accurate crossing from set piece with a goal": "pass",
+}
+
 
 def load_instat_event_data(
     event_data_loc: str, metadata_loc: str
@@ -186,27 +210,27 @@ def _load_event_data(event_data_loc: str, metadata: Metadata) -> pd.DataFrame:
     Returns:
         pd.DataFrame: event data of the match
     """
-    EVENT_AND_OUTCOME_INSTAT_EVENTS = {
-        "Attacking pass accurate": ["pass", 1],
-        "Accurate key pass": ["pass", 1],
-        "Attacking pass inaccurate": ["pass", 0],
-        "Inaccurate key pass": ["pass", 0],
-        "Pass into offside": ["pass", 0],
-        "Successful dribbling": ["dribbling", 1],
-        "Unsuccessful dribbling": ["dribbling", 0],
-        "Dribbling": ["dribbling", MISSING_INT],
-        "Crosses accurate": ["cross", 1],
-        "Accurate crossing from set piece with a shot": ["cross", 1],
-        "Accurate crossing from set piece": ["cross", 1],
-        "Accurate crossing from set piece with a goal": ["cross", 1],
-        "Crosses inaccurate": ["cross", 0],
-        "Cross interception": ["cross", 0],
-        "Inaccurate blocked cross": ["cross", 0],
-        "Inaccurate set-piece cross": ["cross", 0],
-        "Shot on target": ["shot", 1],
-        "Blocked shot": ["shot", 0],
-        "Wide shot": ["shot", 0],
-    }
+    # EVENT_AND_OUTCOME_INSTAT_EVENTS = {
+    #     "Attacking pass accurate": ["pass", 1],
+    #     "Accurate key pass": ["pass", 1],
+    #     "Attacking pass inaccurate": ["pass", 0],
+    #     "Inaccurate key pass": ["pass", 0],
+    #     "Pass into offside": ["pass", 0],
+    #     "Successful dribbling": ["dribbling", 1],
+    #     "Unsuccessful dribbling": ["dribbling", 0],
+    #     "Dribbling": ["dribbling", MISSING_INT],
+    #     "Crosses accurate": ["cross", 1],
+    #     "Accurate crossing from set piece with a shot": ["cross", 1],
+    #     "Accurate crossing from set piece": ["cross", 1],
+    #     "Accurate crossing from set piece with a goal": ["cross", 1],
+    #     "Crosses inaccurate": ["cross", 0],
+    #     "Cross interception": ["cross", 0],
+    #     "Inaccurate blocked cross": ["cross", 0],
+    #     "Inaccurate set-piece cross": ["cross", 0],
+    #     "Shot on target": ["shot", 1],
+    #     "Blocked shot": ["shot", 0],
+    #     "Wide shot": ["shot", 0],
+    # }
 
     with open(event_data_loc, "r") as f:
         data = f.read()
@@ -216,7 +240,8 @@ def _load_event_data(event_data_loc: str, metadata: Metadata) -> pd.DataFrame:
     result_dict = {
         "event_id": [],
         "type_id": [],
-        "event": [],
+        "instat_event": [],
+        "databallpy_event": [],
         "period_id": [],
         "minutes": [],
         "seconds": [],
@@ -228,7 +253,6 @@ def _load_event_data(event_data_loc: str, metadata: Metadata) -> pd.DataFrame:
         "end_x": [],
         "end_y": [],
         "datetime": [],
-        "instat_event": [],
     }
 
     start_time_period = {
@@ -247,15 +271,17 @@ def _load_event_data(event_data_loc: str, metadata: Metadata) -> pd.DataFrame:
         if not event["action_id"].startswith(("16", "15")):
             result_dict["event_id"].append(int(event["id"]))
             result_dict["type_id"].append(int(event["action_id"]))
-            if str(event["action_name"]) in EVENT_AND_OUTCOME_INSTAT_EVENTS.keys():
-                event_name, outcome = EVENT_AND_OUTCOME_INSTAT_EVENTS[
-                    str(event["action_name"])
-                ]
-                result_dict["event"].append(event_name)
-                result_dict["outcome"].append(outcome)
-            else:
-                result_dict["event"].append(str(event["action_name"]))
-                result_dict["outcome"].append(np.nan)
+
+            # if str(event["action_name"]) in EVENT_AND_OUTCOME_INSTAT_EVENTS.keys():
+            #     event_name, outcome = EVENT_AND_OUTCOME_INSTAT_EVENTS[
+            #         str(event["action_name"])
+            #     ]
+            #     result_dict["event"].append(event_name)
+            #     result_dict["outcome"].append(outcome)
+            # else:
+            result_dict["instat_event"].append(str(event["action_name"]))
+            result_dict["outcome"].append(np.nan)
+
             result_dict["period_id"].append(int(event["half"]))
             result_dict["minutes"].append(float(event["second"]) // 60)
             result_dict["seconds"].append(float(event["second"]) % 60)
@@ -283,12 +309,17 @@ def _load_event_data(event_data_loc: str, metadata: Metadata) -> pd.DataFrame:
                 start_time_period[int(event["half"])]
                 + dt.timedelta(milliseconds=float(event["second"]) * 1000)
             )
-            result_dict["instat_event"].append(str(event["action_name"]))
+
+    result_dict["databallpy_event"] = [np.nan] * len(result_dict["event_id"])
 
     event_data = pd.DataFrame(result_dict)
-    start_events = ["pass", "shot", "Goal"]
+    event_data["databallpy_event"] = event_data["instat_event"].map(
+        instat_databallpy_map
+    )
+
+    start_events = ["pass", "shot"]
     x_start, y_start = (
-        event_data[event_data["event"].isin(start_events)]
+        event_data[event_data["databallpy_event"].isin(start_events)]
         .reset_index()
         .loc[0, ["start_x", "start_y"]]
     )

databallpy/load_data/event_data/metrica_event_data.py

@@ -20,6 +20,12 @@
 )
 from databallpy.utils.utils import _to_float, _to_int
 
+metrica_databallpy_map = {
+    "pass": "pass",
+    "carry": "dribble",
+    "shot": "shot",
+}
+
 
 def load_metrica_event_data(
     event_data_loc: str, metadata_loc: str
@@ -78,6 +84,7 @@ def load_metrica_event_data(
         dt.timedelta(milliseconds=(x - first_frame) / frame_rate * 1000)
         for x in event_data["td_frame"]
     ]
+
     # no idea about time zone since we have no real data, so just assume utc
     event_data["datetime"] = [
         pd.to_datetime(start_time, utc=True) + x for x in rel_timedelta
@@ -124,7 +131,8 @@ def _get_event_data(event_data_loc: Union[str, io.StringIO]) -> pd.DataFrame:
     result_dict = {
         "event_id": [],
         "type_id": [],
-        "event": [],
+        "metrica_event": [],
+        "databallpy_event": [],
         "period_id": [],
         "minutes": [],
         "seconds": [],
@@ -141,18 +149,57 @@ def _get_event_data(event_data_loc: Union[str, io.StringIO]) -> pd.DataFrame:
         "td_frame": [],
     }
 
+    check_outcome_last_event = False
+
+    in_posession_events = ["pass", "carry", "recovery", "shot"]
+    out_of_posession_events = ["fault received", "ball out", "ball lost"]
+
     for event in events_dict["data"]:
         result_dict["event_id"].append(event["index"])
         result_dict["type_id"].append(event["type"]["id"])
         event_name = event["type"]["name"].lower()
-        result_dict["event"].append(event_name)
+        result_dict["metrica_event"].append(event_name)
         result_dict["period_id"].append(event["period"])
         result_dict["minutes"].append(_to_int((event["start"]["time"] // 60)))
         result_dict["seconds"].append(_to_float(event["start"]["time"] % 60))
         result_dict["player_id"].append(_to_int(event["from"]["id"][1:]))
         result_dict["player_name"].append(event["from"]["name"])
+
+        # set outcome for pass or carry events
+        if check_outcome_last_event:
+            if (
+                event_name in out_of_posession_events
+                and result_dict["team_id"][-1] == event["team"]["id"]
+            ) or (
+                event_name in in_posession_events
+                and result_dict["team_id"][-1] != event["team"]["id"]
+            ):
+
+                result_dict["outcome"][-1] = 0
+            else:
+                result_dict["outcome"][-1] = 1
+            check_outcome_last_event = False
+
+        # set outcome for shot events
+        if event_name == "shot":
+            if isinstance(event["subtypes"], list):
+                outcome = 0
+                for sub in event["subtypes"]:
+                    if sub["name"] == "GOAL":
+                        outcome = 1
+                        break
+            else:
+                subtypes = event["subtypes"]
+                outcome = 1 if subtypes["name"] == "GOAL" else 0
+            result_dict["outcome"].append(outcome)
+        else:
+            result_dict["outcome"].append(np.nan)
+
+        # Check if outcome needs to be set based on next event
+        if event_name in ["pass", "carry"]:
+            check_outcome_last_event = True
+
         result_dict["team_id"].append(event["team"]["id"])
-        result_dict["outcome"].append(np.nan)
         result_dict["start_x"].append(_to_float(event["start"]["x"]))
         result_dict["start_y"].append(_to_float(event["start"]["y"]))
         if event["to"] is not None:
@@ -165,5 +212,7 @@ def _get_event_data(event_data_loc: Union[str, io.StringIO]) -> pd.DataFrame:
         result_dict["end_y"].append(_to_float(event["end"]["y"]))
         result_dict["td_frame"].append(event["start"]["frame"])
 
+    result_dict["databallpy_event"] = [np.nan] * len(result_dict["event_id"])
     events = pd.DataFrame(result_dict)
+    events["databallpy_event"] = events["metrica_event"].map(metrica_databallpy_map)
     return events

databallpy/load_data/event_data/opta.py

@@ -86,6 +86,16 @@
     77: "player off pitch",
 }
 
+opta_to_databallpy_map = {
+    "pass": "pass",
+    "take on": "dribble",
+    "offside pass": "pass",
+    "miss": "shot",
+    "post": "shot",
+    "attempt saved": "shot",
+    "goal": "shot",
+}
+
 
 def load_opta_event_data(
     f7_loc: str, f24_loc: str, pitch_dimensions: list = [106.0, 68.0]
@@ -323,14 +333,15 @@ def _load_event_data(f24_loc: str, country: str) -> pd.DataFrame:
         pd.DataFrame: all events of the match in a pd dataframe
     """
 
-    file = open(f24_loc, "r")
-    lines = file.read()
+    with open(f24_loc, "r") as file:
+        lines = file.read()
     soup = BeautifulSoup(lines, "xml")
 
     result_dict = {
         "event_id": [],
         "type_id": [],
-        "event": [],
+        "opta_event": [],
+        "databallpy_event": [],
         "period_id": [],
         "minutes": [],
         "seconds": [],
@@ -363,7 +374,7 @@ def _load_event_data(f24_loc: str, country: str) -> pd.DataFrame:
             # Unknown event
             event_name = None
 
-        result_dict["event"].append(event_name)
+        result_dict["opta_event"].append(event_name)
         result_dict["period_id"].append(int(event.attrs["period_id"]))
         result_dict["minutes"].append(int(event.attrs["min"]))
         result_dict["seconds"].append(int(event.attrs["sec"]))
@@ -381,7 +392,10 @@ def _load_event_data(f24_loc: str, country: str) -> pd.DataFrame:
             pd.to_datetime(event.attrs["timestamp"], utc=True)
         )
 
-    file.close()
+    result_dict["databallpy_event"] = [np.nan] * len(result_dict["event_id"])
     event_data = pd.DataFrame(result_dict)
+    event_data["databallpy_event"] = event_data["opta_event"].map(
+        opta_to_databallpy_map
+    )
     event_data["datetime"] = utc_to_local_datetime(event_data["datetime"], country)
     return event_data

databallpy/match.py

@@ -318,7 +318,7 @@ def check_inputs_match_object(match: Match):
     if len(match.event_data) > 0:
         for col in [
             "event_id",
-            "event",
+            "databallpy_event",
             "period_id",
             "team_id",
             "player_id",