impectPy 2.4.4__py3-none-any.whl

impectPy/player_profile_scores.py

@@ -0,0 +1,260 @@
+# load packages
+import pandas as pd
+import requests
+from impectPy.helpers import RateLimitedAPI, unnest_mappings_df
+from .iterations import getIterationsFromHost
+
+# define the allowed positions
+allowed_positions = [
+  "GOALKEEPER",
+  "LEFT_WINGBACK_DEFENDER",
+  "RIGHT_WINGBACK_DEFENDER",
+  "CENTRAL_DEFENDER",
+  "DEFENSE_MIDFIELD",
+  "CENTRAL_MIDFIELD",
+  "ATTACKING_MIDFIELD",
+  "LEFT_WINGER",
+  "RIGHT_WINGER",
+  "CENTER_FORWARD"
+]
+
+######
+#
+# This function returns a pandas dataframe that contains all profile scores
+# for a given iteration and a given set of positions per player
+#
+######
+
+
+def getPlayerProfileScores(
+        iteration: int, positions: list, token: str, session: requests.Session = requests.Session()
+) -> pd.DataFrame:
+
+    # create an instance of RateLimitedAPI
+    connection = RateLimitedAPI(session)
+
+    # construct header with access token
+    connection.session.headers.update({"Authorization": f"Bearer {token}"})
+
+    return getPlayerProfileScoresFromHost(iteration, positions, connection, "https://api.impect.com")
+
+def getPlayerProfileScoresFromHost(
+        iteration: int, positions: list, connection: RateLimitedAPI, host: str
+) -> pd.DataFrame:
+
+    # check input for iteration argument
+    if not isinstance(iteration, int):
+        raise Exception("Input for iteration argument must be an integer")
+
+    # check input for positions argument
+    if not isinstance(positions, list):
+        raise Exception("Input for positions argument must be a list")
+
+    # check if the input positions are valid
+    invalid_positions = [position for position in positions if position not in allowed_positions]
+    if len(invalid_positions) > 0:
+        raise Exception(
+            f"Invalid position(s): {', '.join(invalid_positions)}."
+            f"\nChoose one or more of: {', '.join(allowed_positions)}"
+        )
+
+    # get squads
+    squads = connection.make_api_request_limited(
+        url=f"{host}/v5/customerapi/iterations/{iteration}/squads",
+        method="GET"
+    ).process_response(
+        endpoint="Squads"
+    )
+
+    # get squadIds
+    squad_ids = squads[squads.access].id.to_list()
+
+    # compile position string
+    position_string = ",".join(positions)
+
+    # get player profile scores per squad
+    profile_scores_raw = pd.concat(
+        map(lambda squadId: connection.make_api_request_limited(
+            url=f"{host}/v5/customerapi/iterations/{iteration}/"
+                f"squads/{squadId}/positions/{position_string}/player-profile-scores",
+            method="GET"
+        ).process_response(
+            endpoint="PlayerIterationScores",
+            raise_exception=False
+        ).assign(
+            iterationId=iteration,
+            squadId=squadId,
+            positions=position_string
+        ),
+            squad_ids),
+        ignore_index=True)
+
+    # raise exception if no player played at given positions in entire iteration
+    if len(profile_scores_raw) == 0:
+        raise Exception(f"No players played at given position in iteration {iteration}.")
+
+    # print squads without players at given position
+    error_list = [str(squadId) for squadId in squad_ids if squadId not in profile_scores_raw.squadId.to_list()]
+    if len(error_list) > 0:
+        print(f"No players played at positions {positions} for iteration {iteration} for following squads:\n\t{', '.join(error_list)}")
+
+    # get players
+    players = connection.make_api_request_limited(
+        url=f"{host}/v5/customerapi/iterations/{iteration}/players",
+        method="GET"
+    ).process_response(
+        endpoint="Players"
+    )[["id", "commonname", "firstname", "lastname", "birthdate", "birthplace", "leg", "countryIds", "idMappings"]]
+
+    # only keep first country id for each player
+    country_series = players["countryIds"].explode().groupby(level=0).first()
+    players["countryIds"] = players.index.to_series().map(country_series).astype("float").astype("Int64")
+    players = players.rename(columns={"countryIds": "countryId"})
+
+    # unnest mappings
+    players = unnest_mappings_df(players, "idMappings").drop(["idMappings"], axis=1).drop_duplicates()
+
+    # get scores
+    scores = connection.make_api_request_limited(
+        url=f"{host}/v5/customerapi/player-profiles",
+        method="GET"
+    ).process_response(
+        endpoint="playerProfiles"
+    )[["name"]]
+
+    # get iterations
+    iterations = getIterationsFromHost(connection=connection, host=host)
+
+    # get country data
+    countries = connection.make_api_request_limited(
+        url=f"{host}/v5/customerapi/countries",
+        method="GET"
+    ).process_response(
+        endpoint="KPIs"
+    )
+
+    # unnest scorings
+    profile_scores = profile_scores_raw.explode("profileScores").reset_index(drop=True)
+
+    # unnest dictionary in kpis column
+    profile_scores = pd.concat(
+        [profile_scores.drop(["profileScores"], axis=1), pd.json_normalize(profile_scores["profileScores"])],
+        axis=1
+    )
+
+    # merge with player scores to ensure all kpis are present
+    profile_scores = profile_scores.merge(
+        scores,
+        left_on="profileName",
+        right_on="name",
+        how="outer",
+        suffixes=("", "_right")
+    )
+
+    # get matchShares
+    match_shares = profile_scores[
+        ["iterationId", "squadId", "playerId", "positions", "playDuration", "matchShare"]].drop_duplicates()
+
+    # fill missing values in the "name" column with a default value to ensure players without scorings don't get lost
+    if len(profile_scores["name"][profile_scores["name"].isnull()]) > 0:
+        profile_scores["name"] = profile_scores["name"].fillna("-1")
+
+    # pivot kpi values
+    profile_scores = pd.pivot_table(
+        profile_scores,
+        values="value",
+        index=["iterationId", "squadId", "playerId", "positions"],
+        columns="name",
+        aggfunc="sum",
+        fill_value=0,
+        dropna=False
+    ).reset_index()
+
+    # drop "-1" column
+    if "-1" in profile_scores.columns:
+        profile_scores.drop(["-1"], inplace=True, axis=1)
+
+    # merge with playDuration and matchShare
+    profile_scores = profile_scores.merge(
+        match_shares,
+        left_on=["iterationId", "squadId", "playerId", "positions"],
+        right_on=["iterationId", "squadId", "playerId", "positions"],
+        how="inner",
+        suffixes=("", "_right")
+    )
+    # merge with other data
+    profile_scores = profile_scores.merge(
+        iterations[["id", "competitionName", "season"]],
+        left_on="iterationId",
+        right_on="id",
+        how="left",
+        suffixes=("", "_right")
+    ).merge(
+        squads[["id", "name"]].rename(
+            columns={"id": "squadId", "name": "squadName"}
+        ),
+        left_on="squadId",
+        right_on="squadId",
+        how="left",
+        suffixes=("", "_right")
+    ).merge(
+        players[[
+            "id", "wyscoutId", "heimSpielId", "skillCornerId", "commonname",
+            "firstname", "lastname", "birthdate", "birthplace", "countryId", "leg"
+        ]].rename(
+            columns={"commonname": "playerName"}
+        ),
+        left_on="playerId",
+        right_on="id",
+        how="left",
+        suffixes=("", "_right")
+    ).merge(
+        countries.rename(columns={"fifaName": "playerCountry"}),
+        left_on="countryId",
+        right_on="id",
+        how="left",
+        suffixes=("", "_right")
+    )
+
+    # remove NA rows
+    profile_scores = profile_scores[profile_scores.iterationId.notnull()]
+
+    # fix column types
+    profile_scores["squadId"] = profile_scores["squadId"].astype("Int64")
+    profile_scores["playerId"] = profile_scores["playerId"].astype("Int64")
+    profile_scores["iterationId"] = profile_scores["iterationId"].astype("Int64")
+    profile_scores["wyscoutId"] = profile_scores["wyscoutId"].astype("Int64")
+    profile_scores["heimSpielId"] = profile_scores["heimSpielId"].astype("Int64")
+    profile_scores["skillCornerId"] = profile_scores["skillCornerId"].astype("Int64")
+
+    # define column order
+    order = [
+        "iterationId",
+        "competitionName",
+        "season",
+        "squadId",
+        "squadName",
+        "playerId",
+        "wyscoutId",
+        "heimSpielId",
+        "skillCornerId",
+        "playerName",
+        "firstname",
+        "lastname",
+        "birthdate",
+        "birthplace",
+        "playerCountry",
+        "leg",
+        "positions",
+        "matchShare",
+        "playDuration"
+    ]
+
+    # add kpiNames to order
+    order = order + scores.name.to_list()
+
+    # select columns
+    profile_scores = profile_scores[order]
+
+    # return result
+    return profile_scores