train_model.py

import model
import scrape_data
import os
import pandas as pd
import numpy as np


def get_team_stats(team: list) -> list:
    stats = []
    for player in team:
        stats.extend(scrape_data.df_to_stats(scrape_data.scrape_data_from_profile(player))) # fix to get average stats from inner function
    return stats


def create_data():
    if os.path.exists((os.path.join(os.getcwd(), "training_data.csv"))):
        print("data already exists")
        return
    columns = [str(i) for i in range(7*5 + 1)]
    df = pd.DataFrame(columns=columns)
    player_data = scrape_data.get_players(2, 5)
    for count, i in enumerate(player_data):
        team = scrape_data.get_last_game_from_profile(i)
        team_stats = get_team_stats(team[:-1])
        team_df = pd.DataFrame(team_stats).T
        if len(team_stats) != 0:
            team_df[36] = team[-1]
            team_df.columns = df.columns
            df = pd.concat([df, team_df], ignore_index=True)
        if count % 5 == 0:
            print(count)
        if count % 50 == 0:
            df.to_csv("training_data.csv")
    df.to_csv("training_data.csv")
    print("complete")

def start_training():
    df = pd.read_csv('sample_data.csv')
    X = df.iloc[:, 1:-1].values
    y = df.iloc[:, -1].values
    leagueAi = model.LeagueModel()
    model.train(leagueAi, X, y, 10000, True)

if __name__ == "__main__":
    create_data()
    #start_training()