data = df[start_p:stop_p].astype('float32') # data['Day'] = data.index.dayofyear #add day data = data.interpolate(limit=30000000, limit_direction='both').astype( 'float32') #interpolate neighbor first, for rest NA fill with mean() data[target] = data[target].shift(-out_t_step) data.dropna(inplace=True) return data #--------------------------------------------------------# loading = instant_data() df, mode = loading.hourly_instant(), 'hour' # df,mode = loading.daily_instant(),'day' st = 'CPY012' target, start_p, stop_p, host_path = station_sel(st, mode) if mode == 'hour': n_past, n_future = 24 * 6, 72 elif mode == 'day': n_past, n_future = 60, 30 n_pca = 4 split_date = '2016-11-01' #----------------------------------------# save_path = host_path + 'ML_svr/' if not os.path.exists(save_path): os.makedirs(save_path) ########################################### def call_data(): loading = instant_data() df, mode = loading.hourly_instant(), 'hour'
import matplotlib.pyplot as plt from tqdm import tqdm from DLtools.Data import instant_data, check_specific_col, station_sel from DLtools.MachineLearning import tsplot def scope_data(data): global start_p, stop_p data = data[start_p:stop_p] # data = del_less_col(data) return data ############################## st, mode = 'CPY012', 'day' target, start_p, stop_p, _ = station_sel(st, mode) ############################## loading = instant_data() df_r = scope_data(loading.df_r) df_w = scope_data(loading.df_w) df_wet = scope_data(loading.df_wet) df_dam = scope_data(loading.df_d) # df_day=loading.daily_instant() # df_hour =loading.hourly_instant() df_solar = df_wet[check_specific_col(df_wet, 'solar')] df_rain1h = df_wet[check_specific_col(df_wet, 'rain1h')] df_temp = df_wet[check_specific_col(df_wet, 'temp')] df_press = df_wet[check_specific_col(df_wet, 'press')] df_humid = df_wet[check_specific_col(df_wet, 'humid')]