def update_timeSeries_graph(json_data, var):
try:
# call dat
df = pd.read_json(json_data)
colname = [col for col in df.columns if var in col]
df = df[colname]
fig = get_ts_plot(df, yaxis_label='')
return fig
except:
return []
def update_corr_graph(json_data, step):
try:
df = pd.read_json(json_data).copy()
# current vs power corr
for i in range(1, 4):
df[f'Corr{i}'] = df[f'S{i}'].rolling(step).corr(
df[f'I{i}'].rolling(step))
df_corr = df[[col for col in df.columns if 'Corr' in col]].copy()
fig_corr = get_ts_plot(df_corr, yaxis_label=' ', hover_mode='x')
return fig_corr
except:
return []
def update_gas_graph(json_data):
try:
df_gas = pd.read_json(json_data)
df_gas = df_gas[['CH4', 'C2H2', 'C2H4', 'C2H6', 'H2']]
# plotting
fig = get_ts_plot(df_gas,
yaxis_label='ppm',
convert_x=False,
drag_mode='select')
return fig
except:
return []
def update_volt_excurs_graph(json_data, ma_win):
try:
# call data
df_trans = pd.read_json(json_data).copy()
# compute excursion percentage
df_1 = get_excur_percent(df_trans, ma_win, channel='V')
# plotting
fig = get_ts_plot(df_1, yaxis_label='% Change (from MA)')
return fig
except:
return []
def update_pf_graph(json_data):
try:
df = pd.read_json(json_data).copy()
# power factor
df_P = df[[col for col in df.columns if 'P' in col]].copy()
df_S = df[[col for col in df.columns if 'S' in col]].copy()
df_pf = df_P.values / df_S.values
df_pf = pd.DataFrame(df_pf,
index=df_P.index,
columns=['Pf1', 'Pf2', 'Pf3'])
fig_pf = get_ts_plot(df_pf, yaxis_label=' ', hover_mode='x')
return fig_pf
except:
return []
def generate_pv_plot(json_data):
try:
datasets = json.loads(json_data)
df_load = pd.read_json(datasets['df_load'])
# df_temp = df_load.dropna()
df_load = df_load[['Power Demand', 'PV Production']]
df_load.index = pd.to_datetime(df_load.index, errors='coerce')
# df_trunc = df_combined.iloc[-60:,:].copy()
df_trunc = df_load[df_load.index > df_load.index[-1] -
timedelta(hours=60)].copy()
fig = get_ts_plot(df_trunc, yaxis_label='Power [KW]')
pv_peak = round(get_pv_peak(df_load['PV Production']), 3)
# evaluate performance
return fig, 'Note: PV is estimated based on GHI, Power Peak (past 60 days) and Real Power.', pv_peak
except:
return [], 'Loading...', 0
def update_average_graph(datetype, json_data, var):
df = pd.read_json(json_data)
df = df[[col for col in df.columns if var in col]]
if datetype == 'Month':
df = df.groupby(df.index.month).mean()
df.index = df.index.map(lambda x: calendar.month_name[x])
elif datetype == 'Hour':
df = df.groupby(df.index.hour).mean()
elif datetype == 'Dai':
df = df.groupby(df.index.day).mean()
else:
df = df.groupby(df.index.dayofweek).mean()
df.index = df.index.map(lambda x: calendar.day_name[x])
fig = get_ts_plot(df, yaxis_label='', convert_x=False, xtick=1)
return fig
def update_normalized_graph(json_data, phase, var):
try:
df = pd.read_json(json_data)
if var != 'All':
df = df[[col for col in df.columns if var in col]]
else:
pass
if phase != 'All':
# filter by phase
df = df[[col for col in df.columns if col[-1] == str(phase)]]
else:
pass
df_zs = (df - df.mean()) / df.std()
fig = get_ts_plot(df_zs, yaxis_label='z-score')
return fig
except:
return []
def update_tsbox_graph(rows, selectedRows, json_data):
try:
# call data according to selected station
df = pd.read_json(json_data)
df_selected = pd.DataFrame(rows).loc[selectedRows]
df_selected['target'] = df_selected['Station'] + '__' + df_selected[
'Variable (phase)']
df = df[[
col for col in df.columns if col in df_selected['target'].unique()
]]
# plotting
fig_ts = get_ts_plot(df,
yaxis_label=' ',
convert_x=False,
xaxis_label=' ')
fig_box = get_box_plot(df, y_label=' ')
return fig_ts, fig_box
except:
return [], []
def generate_real_time_plot(json_data):
try:
datasets = json.loads(json_data)
df_combined = pd.read_json(datasets['df_combined'])
df_combined.index = pd.to_datetime(df_combined.index, errors='coerce')
df_trunc = df_combined[df_combined.index > df_combined.index[-1] -
timedelta(hours=60)].copy()
df_t = df_combined.dropna()
fig = get_ts_plot(df_trunc,
yaxis_label='Active Power [kW]',
hover_mode='closest')
r2 = r2_score(df_t[df_t.columns[0]], df_t[df_t.columns[1]])
rmse = np.sqrt(
mean_squared_error(df_t[df_t.columns[0]], df_t[df_t.columns[1]]))
mae = mean_absolute_error(df_t[df_t.columns[0]], df_t[df_t.columns[1]])
fc_p = 'Performance: r2 = {}______RMSE = {}______MAE = {}'.format(
round(r2, 3), round(rmse, 3), round(mae, 3))
return fig, fc_p
except:
return [], 'Loading...'
def update_dev_graph(json_data, var):
try:
# call data
df = pd.read_json(json_data)
df = df[[col for col in df.columns if var in col]]
# comptue MAD
df_mad = df.mad(axis=1)
alpha = 1 - 0.1
idx = get_sh_esd(df_mad, alpha=alpha)
# get outlier
if idx != None:
df_out = df_mad[idx]
fig = get_ts_plot_outlier(df_mad, df_out, yaxis_label='MAD')
else:
fig = get_ts_plot(
df_mad,
title='Anamoly Detection of Unbalances in {}'.format(var),
yaxis_label='MAD',
)
return fig
except:
return []
def update_co_ratio_graph(json_data):
try:
#ratio
df_gas = pd.read_json(json_data)
# CO2/CO figure
df_r1 = pd.DataFrame({
'CO2/CO': df_gas['CO2'] / df_gas['CO'],
'th': 3.0,
})
df_r1.replace([np.inf, np.nan], 0, inplace=True)
fig_co = get_ts_plot(df_r1,
yaxis_label='',
convert_x=False,
legend_bool=False)
# C2H2/H2 ratio plot
df_r2 = pd.DataFrame({
'C2H2/H2': df_gas['C2H2'] / df_gas['H2'],
'low_th': 2.0,
'high_th': 3.0,
})
df_r2.replace([np.inf, np.nan], 0, inplace=True)
fig_ch = get_ts_plot(df_r2,
yaxis_label='',
convert_x=False,
legend_bool=True)
# O2/N2 ratio plot
df_r3 = pd.DataFrame({'O2/N2': df_gas['O2'] / df_gas['N2']})
df_r3.replace([np.inf, np.nan], 0, inplace=True)
fig_on = get_ts_plot(df_r3,
yaxis_label='',
convert_x=False,
legend_bool=False)
# error log details
report = [
html.H2('Analytic Result for DGA', className='test-center'),
]
# for CO
for i in df_r1[df_r1['CO2/CO'] < 3].index.tolist():
report.append(
html.Div('[{}]: Paper Fault (CO2/CO ratio)'.format(i), ))
# for C2H2/H2
for i in df_r2[(df_r2['C2H2/H2'] > 2)
& (df_r2['C2H2/H2'] < 3)].index.tolist():
print(i)
report.append(
html.Div(
'[{}]: LTC Contamination (C2H2/H2 ratio)'.format(i), ))
# for O2/N2
df_r3['decreasing'] = np.nan
for i in range(len(df_r3)):
x = df_r3.iloc[i:i + 4, 0]
if ((x.empty) | (x.mean() == 0)):
df_r3.iloc[i, -1] = np.nan
else:
df_r3.iloc[i, -1] = x.is_monotonic_decreasing
# for C2H2/H2
for i in df_r3[df_r3['decreasing'].diff() == 1].index.tolist():
print(i)
report.append(
html.Div(
'[{}]: Excessive Heating (Decreasing O2/N2 ratio)'.format(
i), ))
return fig_co, fig_ch, fig_on, report
except:
return [], [], [], []