from datasets import DatasetBuilder, DatasetInfo, Features, Value, SplitGenerator import pandas as pd class Euro2020Dataset(DatasetBuilder): VERSION = "1.0.0" def _info(self): return DatasetInfo( description="Your dataset description here.", features=Features({ "HomeTeamName": Value("string"), "AwayTeamName": Value("string"), "DateandTimeCET": Value("string"), "MatchID": Value("int64"), "RoundName": Value("string"), "Stage": Value("string"), "MatchDay": Value("int64"), "Session": Value("int64"), "MatchMinute": Value("int64"), "InjuryTime": Value("int64"), "NumberofPhases": Value("int64"), "Phase": Value("int64"), "ScoreHome": Value("int64"), "ScoreAway": Value("int64"), "MatchStatus": Value("string"), "StadiumID": Value("int64"), "NumberofMatchesRefereedPostMatch": Value("int64"), "TotalNumberofMatchesRefereed": Value("int64"), "NumberofMatchesRefereedinGroupStage": Value("int64"), "NumberofMatchesRefereedinKnockoutStage": Value("int64"), "AssistantRefereeWebName": Value("string"), "Humidity": Value("int64"), "Temperature": Value("int64"), "WindSpeed": Value("int64"), "MatchEvent": Features({ "1-First Half": Features({ "Event": Value("string"), "Minute": Value("int"), "Phase": Value("int"), "InjuryMinute": Value("int"), "TeamFromID": Value("float"), "TeamToID": Value("float"), "PlayerFromID": Value("float"), "PlayerToID": Value("float"), "Time": Value("string"), "MatchEventAttribute": Value("float"), }), "2-Second Half": Features({ "Event": Value("string"), "Minute": Value("int"), "Phase": Value("int"), "InjuryMinute": Value("int"), "TeamFromID": Value("float"), "TeamToID": Value("float"), "PlayerFromID": Value("float"), "PlayerToID": Value("float"), "Time": Value("string"), "MatchEventAttribute": Value("float"), }), "3-Extra Time First Half": Features({ "Event": Value("string"), "Minute": Value("int"), "Phase": Value("int"), "InjuryMinute": Value("int"), "TeamFromID": Value("float"), "TeamToID": Value("float"), "PlayerFromID": Value("float"), "PlayerToID": Value("float"), "Time": Value("string"), "MatchEventAttribute": Value("float"), }), "4-Extra Time Second Half": Features({ "Event": Value("string"), "Minute": Value("int"), "Phase": Value("int"), "InjuryMinute": Value("int"), "TeamFromID": Value("float"), "TeamToID": Value("float"), "PlayerFromID": Value("float"), "PlayerToID": Value("float"), "Time": Value("string"), "MatchEventAttribute": Value("float"), }), "5-Penalty Shootout": Features({ "Event": Value("string"), "Minute": Value("int"), "Phase": Value("int"), "InjuryMinute": Value("int"), "TeamFromID": Value("float"), "TeamToID": Value("float"), "PlayerFromID": Value("float"), "PlayerToID": Value("float"), "Time": Value("string"), "MatchEventAttribute": Value("float"), }), }), "TeamLineUps": Value("dict"), "TeamStats": Value("dict"), "PlayerStats": Value("dict"), "PlayerPreMatchInfo": Value("dict"), }), supervised_keys=None, homepage="Optional dataset homepage", license= "", citation="Optional citation" ) def _split_generators(self, dl_manager): # Since the final processed dataset is a single CSV file that combines the content from numerous rows to only 51 rows, # with each row representing the information of each game, I wil just use a single 'train' split and no need for test / validation. return [ SplitGenerator( name="train", gen_kwargs={"filepath": "ForzaJuve1/UEFA_Euro_2020_Data"} ), ] #def _generate_examples(self, filepath): #Here we read the CSV file and yield examples #data = pd.read_csv(filepath) #for idx, row in data.iterrows(): #yield idx, { #"PlayerID": row["PlayerID"], #"PlayerName": row["PlayerName"], #"PlayerSurname": row["PlayerSurname"], #"IsGoalkeeper": row["IsGoalkeeper"], #"PlayedTime": row["PlayedTime"], #"StatsID": row["StatsID"], #"StatsName": row["StatsName"], #"Value": row["Value"], #"Rank": row["Rank"], # Make sure to include all fields defined in _info #}