time_period_cutoffs = (30, 180, 420, 450)
#quotes, trades = get_data('XLE', 2012, 1, 5, bar_width='second')
data = get_more_data('XLE', 2012, 2, 1, days=1, bar_width='second')
hls = [10, 40, 100]
for quotes, trades in data:
features.label_data(quotes, label_hls=hls)
features.add_ema(quotes, halflives=hls)
features.add_dema(quotes, halflives=hls)
features.add_momentum(quotes, halflives=hls)
features.add_log_return_ema(quotes, halflives=hls)
#features.add_trade_momentum(quotes, trades, bar_width='second')
features.add_price_diff(quotes)
features.add_size_diff(quotes)
#features.add_trade_momentum_dema(quotes, trades, halflife=40)
quotes_list, trades_list = zip(*data)
quotes = pd.concat(quotes_list)
trades = pd.concat(trades_list)
feature_names = ['momentum', 'dEMA_10', 'dEMA_40', 'dEMA_100',
'log_returns_10-', 'log_returns_40-', 'log_returns_100-',
'log_returns_std_10-', 'log_returns_std_40-', 'size_diff',
'price_diff', 'VPIN_TIME']
quotes = quotes.fillna(0)
# normalize features
hls = [10, 40, 100]
crossover_hls = (10, 40)
filter_feature_names = []
#filter_feature_names = ['crossover?', 'high_momentum?', 'dEMA_10_sign_change?', 'dEMA_40_sign_change?']
feature_names = []
for data, trades in datas:
features.add_future_log_returns(data, label_hls=hls)
features.add_future_log_returns_rolling(data, windows=(10, 30, 60))
feature_names += features.add_price_dema(data, halflives=hls)
feature_names += features.add_momentum(data, halflives=hls)
feature_names += features.add_log_return_ema(data, halflives=hls)
feature_names += features.add_price_diff(data)
feature_names += features.add_size_diff(data)
#feature_names += features.add_vpin_time(data, window=dt.timedelta(seconds=20))
feature_names += features.add_dema(data, features=['momentum', 'dEMA_10', 'dEMA_40', 'dEMA_100'])
#filter_feature_names += features.add_crossover_filter(data, halflives=crossover_hls)
#filter_feature_names += features.add_high_momentum_filter(data, cutoff=0.1, halflives=(10, 40, 100))
#filter_feature_names += features.add_dema_sign_change_filter(data, halflife=10)
#filter_feature_names += features.add_dema_sign_change_filter(data, halflife=40)
features.add_crossover_filter(data, halflives=crossover_hls)
features.add_high_momentum_filter(data, cutoff=0.1, halflives=(10, 40, 100))
add_dema_sign_change_filter(data, halflife=10)
add_dema_sign_change_filter(data, halflife=40)
feature_names = list(set(feature_names))
data = pd.concat(zip(*datas)[0])
