def create_record(pull_override=None):
fte_df = get_fte(pull_override=pull_override)
trimmed_df = fte_df[['date', 'team1', 'team2', 'score1', 'score2']].copy()
outcome_series = trimmed_df['score1'] > trimmed_df['score2']
outcome_val = outcome_series.values
trimmed_df['outcome'] = outcome_val
valid_team1 = trimmed_df['team1'].values
valid_team2 = trimmed_df['team2'].values
valid_score1 = trimmed_df['score1'].values
valid_score2 = trimmed_df['score2'].values
dates = trimmed_df['date'].values
winners = []
for ii in range(trimmed_df.shape[0]):
if valid_score1[ii] > valid_score2[ii]:
winners.append(valid_team1[ii])
else:
winners.append(valid_team2[ii])
matchup_labels = [
t1 + " " + t2 for t1, t2 in zip(valid_team1, valid_team2)
]
record = pd.DataFrame(data=winners,
index=[dates, matchup_labels],
columns=['winner'])
# for ii in range(trimmed_df.shape[0])
return record
def create_raptor_model(record, pull_override):
fte_df = get_fte(pull_override=pull_override)
dates = fte_df['date'].values
prediction_history = np.zeros(fte_df.shape[0])
for ii in range(fte_df.shape[0]):
date = fte_df['date'].iloc[ii]
prob1 = fte_df['raptor_prob1'].iloc[ii]
prob2 = fte_df['raptor_prob2'].iloc[ii]
team1 = fte_df['team1'].iloc[ii]
team2 = fte_df['team2'].iloc[ii]
if prob1 > prob2:
if record.loc[date, team1 + " " + team2]['winner'] == team1:
prediction_history[ii] = 1
else:
if record.loc[date, team1 + " " + team2]['winner'] == team2:
prediction_history[ii] = 1
return Model(prediction_history, dates, "RAPTOR")