diff --git a/.DS_Store b/.DS_Store index 471757c..fe13f73 100644 Binary files a/.DS_Store and b/.DS_Store differ diff --git a/app/src/flask-server/modele/__pycache__/modele.cpython-39.pyc b/app/src/flask-server/modele/__pycache__/modele.cpython-39.pyc index 70e9874..1b53ba2 100644 Binary files a/app/src/flask-server/modele/__pycache__/modele.cpython-39.pyc and b/app/src/flask-server/modele/__pycache__/modele.cpython-39.pyc differ diff --git a/app/src/flask-server/modele/modele.py b/app/src/flask-server/modele/modele.py index 60891f6..2a2bd25 100644 --- a/app/src/flask-server/modele/modele.py +++ b/app/src/flask-server/modele/modele.py @@ -3,9 +3,9 @@ import pandas as pd from math import sqrt # Funkcja zwraca prawdopodobieństwo zdobycia gola -def LogisticRegression_predict_proba(position_x, position_y, angle, match_minute, Number_Intervening_Opponents, Number_Intervening_Teammates, isFoot, isHead): +def LogisticRegression_predict_proba(position_x, position_y, distance_to_goalM, angle, match_minute, Number_Intervening_Opponents, Number_Intervening_Teammates, isFoot, isHead): - distance_to_goalM = sqrt(( (position_x**2) + (position_y**2))) + # distance_to_goalM = sqrt(( (position_x**2) + (position_y**2))) model = load('regresja_logistyczna.joblib') X_new = pd.DataFrame(columns=['position_x', 'position_y', 'distance_to_goalM', 'angle','match_minute', 'Number_Intervening_Opponents','Number_Intervening_Teammates', 'isFoot', 'isHead']) @@ -14,7 +14,79 @@ def LogisticRegression_predict_proba(position_x, position_y, angle, match_minute #kolejne modele -def xgboost_predict_proba(): +def xgboost_predict_proba(minute, position_name, shot_body_part_name, shot_technique_name, + shot_type_name, shot_first_time, shot_one_on_one, + shot_aerial_won, shot_deflected, shot_open_goal, + shot_follows_dribble, shot_redirect, x1, y1, + number_of_players_opponents, number_of_players_teammates, + angle, distance, x_player_opponent_Goalkeeper, + x_player_opponent_8, x_player_opponent_1, x_player_opponent_2, + x_player_opponent_3, x_player_teammate_1, x_player_opponent_4, + x_player_opponent_5, x_player_opponent_6, x_player_teammate_2, + x_player_opponent_9, x_player_opponent_10, x_player_opponent_11, + x_player_teammate_3, x_player_teammate_4, x_player_teammate_5, + x_player_teammate_6, x_player_teammate_7, x_player_teammate_8, + x_player_teammate_9, x_player_teammate_10, + y_player_opponent_Goalkeeper, y_player_opponent_8, + y_player_opponent_1, y_player_opponent_2, y_player_opponent_3, + y_player_teammate_1, y_player_opponent_4, y_player_opponent_5, + y_player_opponent_6, y_player_teammate_2, y_player_opponent_9, + y_player_opponent_10, y_player_opponent_11, y_player_teammate_3, + y_player_teammate_4, y_player_teammate_5, y_player_teammate_6, + y_player_teammate_7, y_player_teammate_8, y_player_teammate_9, + y_player_teammate_10, x_player_opponent_7, y_player_opponent_7, + x_player_teammate_Goalkeeper, y_player_teammate_Goalkeeper, + shot_kick_off): model = load('xgboost.joblib') + X_new = pd.DataFrame(columns=['minute', 'position_name', 'shot_body_part_name', 'shot_technique_name', + 'shot_type_name', 'shot_first_time', 'shot_one_on_one', + 'shot_aerial_won', 'shot_deflected', 'shot_open_goal', + 'shot_follows_dribble', 'shot_redirect', 'x1', 'y1', + 'number_of_players_opponents', 'number_of_players_teammates', + 'angle', 'distance', 'x_player_opponent_Goalkeeper', + 'x_player_opponent_8', 'x_player_opponent_1', 'x_player_opponent_2', + 'x_player_opponent_3', 'x_player_teammate_1', 'x_player_opponent_4', + 'x_player_opponent_5', 'x_player_opponent_6', 'x_player_teammate_2', + 'x_player_opponent_9', 'x_player_opponent_10', 'x_player_opponent_11', + 'x_player_teammate_3', 'x_player_teammate_4', 'x_player_teammate_5', + 'x_player_teammate_6', 'x_player_teammate_7', 'x_player_teammate_8', + 'x_player_teammate_9', 'x_player_teammate_10', + 'y_player_opponent_Goalkeeper', 'y_player_opponent_8', + 'y_player_opponent_1', 'y_player_opponent_2', 'y_player_opponent_3', + 'y_player_teammate_1', 'y_player_opponent_4', 'y_player_opponent_5', + 'y_player_opponent_6', 'y_player_teammate_2', 'y_player_opponent_9', + 'y_player_opponent_10', 'y_player_opponent_11', 'y_player_teammate_3', + 'y_player_teammate_4', 'y_player_teammate_5', 'y_player_teammate_6', + 'y_player_teammate_7', 'y_player_teammate_8', 'y_player_teammate_9', + 'y_player_teammate_10', 'x_player_opponent_7', 'y_player_opponent_7', + 'x_player_teammate_Goalkeeper', 'y_player_teammate_Goalkeeper', + 'shot_kick_off']) + + X_new.loc[len(X_new.index)] = [minute, position_name, shot_body_part_name, shot_technique_name, + shot_type_name, shot_first_time, shot_one_on_one, + shot_aerial_won, shot_deflected, shot_open_goal, + shot_follows_dribble, shot_redirect, x1, y1, + number_of_players_opponents, number_of_players_teammates, + angle, distance, x_player_opponent_Goalkeeper, + x_player_opponent_8, x_player_opponent_1, x_player_opponent_2, + x_player_opponent_3, x_player_teammate_1, x_player_opponent_4, + x_player_opponent_5, x_player_opponent_6, x_player_teammate_2, + x_player_opponent_9, x_player_opponent_10, x_player_opponent_11, + x_player_teammate_3, x_player_teammate_4, x_player_teammate_5, + x_player_teammate_6, x_player_teammate_7, x_player_teammate_8, + x_player_teammate_9, x_player_teammate_10, + y_player_opponent_Goalkeeper, y_player_opponent_8, + y_player_opponent_1, y_player_opponent_2, y_player_opponent_3, + y_player_teammate_1, y_player_opponent_4, y_player_opponent_5, + y_player_opponent_6, y_player_teammate_2, y_player_opponent_9, + y_player_opponent_10, y_player_opponent_11, y_player_teammate_3, + y_player_teammate_4, y_player_teammate_5, y_player_teammate_6, + y_player_teammate_7, y_player_teammate_8, y_player_teammate_9, + y_player_teammate_10, x_player_opponent_7, y_player_opponent_7, + x_player_teammate_Goalkeeper, y_player_teammate_Goalkeeper, + shot_kick_off] + return model.predict_proba(X_new)[0][1].round(2) + + diff --git a/app/src/flask-server/server.py b/app/src/flask-server/server.py index c8beb7c..0e07f2c 100644 --- a/app/src/flask-server/server.py +++ b/app/src/flask-server/server.py @@ -1,12 +1,26 @@ from flask import Flask, request, jsonify from flask_cors import CORS -from modele.modele import LogisticRegression_predict_proba +from modele.modele import LogisticRegression_predict_proba, xgboost_predict_proba +import math app = Flask(__name__) CORS(app) app.config['CORS_HEADERS'] = 'Content-Type' +def loc2angle(x, y): + rads = math.atan(7.32 * x / (x**2 + (y - 34)**2 - (7.32/2)**2)) + rads = math.pi + rads if rads < 0 else rads + deg = math.degrees(rads) + return deg + +def loc2distance(x, y): + return math.sqrt(x**2 + (y - 34)**2) + +def loc2locdistance(x1, y1, x2, y2): + return math.sqrt((x1 - x2)**2 + (y1 - y2)**2) + + # model Api # @app.route("/members") # def members(): @@ -16,6 +30,51 @@ app.config['CORS_HEADERS'] = 'Content-Type' # @app.route("/LRegresion&") @app.route("/get_model", methods = ['GET']) +# def get_model(): + +# # x = int(x[0:2]) +# # y = int(y[0:2]) +# x = request.args.get('x', type=float) +# y = request.args.get('y', type=float) +# ## change model on xgboost +# ## add angle, match minutes and number of players + +# angle = loc2angle(x = x, y = y) +# dist = loc2distance(x = x, y = y) + +# if y is None and x is None: +# return jsonify({"error": "Brak wymaganych parametrów"}), 400 + +# response = xgboost_predict_proba(minute = 20, position_name, shot_body_part_name, shot_technique_name, +# shot_type_name, shot_first_time, shot_one_on_one, +# shot_aerial_won, shot_deflected, shot_open_goal, +# shot_follows_dribble, shot_redirect, x1 = x, y1 = y, +# number_of_players_opponents, number_of_players_teammates, +# angle = angle, distance = dist, x_player_opponent_Goalkeeper, +# x_player_opponent_8, x_player_opponent_1, x_player_opponent_2, +# x_player_opponent_3, x_player_teammate_1, x_player_opponent_4, +# x_player_opponent_5, x_player_opponent_6, x_player_teammate_2, +# x_player_opponent_9, x_player_opponent_10, x_player_opponent_11, +# x_player_teammate_3, x_player_teammate_4, x_player_teammate_5, +# x_player_teammate_6, x_player_teammate_7, x_player_teammate_8, +# x_player_teammate_9, x_player_teammate_10, +# y_player_opponent_Goalkeeper, y_player_opponent_8, +# y_player_opponent_1, y_player_opponent_2, y_player_opponent_3, +# y_player_teammate_1, y_player_opponent_4, y_player_opponent_5, +# y_player_opponent_6, y_player_teammate_2, y_player_opponent_9, +# y_player_opponent_10, y_player_opponent_11, y_player_teammate_3, +# y_player_teammate_4, y_player_teammate_5, y_player_teammate_6, +# y_player_teammate_7, y_player_teammate_8, y_player_teammate_9, +# y_player_teammate_10, x_player_opponent_7, y_player_opponent_7, +# x_player_teammate_Goalkeeper, y_player_teammate_Goalkeeper, +# shot_kick_off) +# #print(x) +# #print(y) +# #print(response) +# res = str(response) +# #return {"response":res} +# return jsonify({"response":res}) + def get_model(): # x = int(x[0:2]) @@ -25,12 +84,16 @@ def get_model(): ## change model on xgboost ## add angle, match minutes and number of players + angle = loc2angle(x = x, y = y) + dist = loc2distance(x = x, y = y) + if y is None and x is None: return jsonify({"error": "Brak wymaganych parametrów"}), 400 response = LogisticRegression_predict_proba(position_x=x, position_y=y, - angle = 13.67, + distance_to_goalM = dist, + angle = angle, match_minute=13, Number_Intervening_Opponents=3, Number_Intervening_Teammates=0, diff --git a/notebooks/dataCleaning.R b/notebooks/dataCleaning.R index 43c3888..29aa883 100644 --- a/notebooks/dataCleaning.R +++ b/notebooks/dataCleaning.R @@ -114,7 +114,7 @@ loc2angle <- function(x, y) { # distance to goal loc2distance <- function(x, y) { - sqrt(x^2 + y^2) + sqrt(x^2 + (y - 34)^2) } # distance between two points on the pitch @@ -345,5 +345,5 @@ cols <- names(data3_final)[grepl(pattern, names(data3_final))] data_final <- data3_final %>% unnest(all_of(cols)) skimr::skim(data_final) write_csv(data_final, file = "data/final_data.csv") -# df_test <- read.csv("data/final_data.csv", nrows = 100) +df_test <- read.csv("data/final_data.csv", nrows = 1000) ##################### The fourth dataset ##############################