def earthquake(site_id, ts):
"""Insert values to earthquake_events, earthquake_alerts, and
operational_triggers to (re)trigger subsurface alert.
Args:
site_id (int): ID of site to compute earthquake analysis for.
ts (datetime): Timestamp of alert trigger.
"""
# writes to earthquake_events; defaults epicenter to site coordinates, depth to 0, and magnitude to 4.3
sites = eq.get_sites()
earthquake_events = sites.loc[sites.site_id == site_id, ['latitude', 'longitude', 'province']]
earthquake_events.loc[:, 'ts'] = ts
earthquake_events.loc[:, 'magnitude'] = 4.3
earthquake_events.loc[:, 'depth'] = 0
earthquake_events.loc[:, 'critical_distance'] = np.round(eq.get_crit_dist(4.3), decimals=2)
earthquake_events.loc[:, 'issuer'] = 'TOPSSOFTWAREINFRA'
earthquake_events.loc[:, 'processed'] = 1
eq_id = int(db.df_write(data_table = sms.DataTable("earthquake_events", earthquake_events), resource='sensor_data', last_insert=True)[0][0])
# writes to earthquake_alerts
earthquake_alerts = pd.DataFrame({'eq_id': [eq_id], 'site_id': [site_id], 'distance': [0]})
db.df_write(data_table = sms.DataTable("earthquake_alerts", earthquake_alerts), resource='sensor_data')
# writes to operational_triggers
trigger_symbol = mem.get('df_trigger_symbols')
trigger_sym_id = trigger_symbol.loc[(trigger_symbol.trigger_source == 'earthquake') & (trigger_symbol.alert_level == 1), 'trigger_sym_id'].values[0]
operational_trigger = pd.DataFrame({'site_id': [site_id], 'trigger_sym_id': [trigger_sym_id], 'ts': [ts], 'ts_updated': [ts]})
qdb.alert_to_db(operational_trigger, 'operational_triggers')
# details for trigger tech info
earthquake_events.loc[:, 'distance'] = 0
return earthquake_events
def main(end_ts=datetime.now()):
start_ts = pd.to_datetime(end_ts) - timedelta(2)
print(start_ts)
surficial_triggers = qdb.get_surficial_trigger(start_ts, end_ts)
if len(surficial_triggers) == 0:
qdb.print_out("No surficial trigger to process")
for index, surficial in surficial_triggers.iterrows():
ts_updated = surficial['ts_updated']
public_ts_start = round_data_ts(ts_updated)
alert_level = surficial['alert_level']
alert_symbol = surficial['alert_symbol']
alert_status = surficial['alert_status']
site_id = surficial['site_id']
site_code = surficial['site_code']
if (alert_symbol == 'lt'):
if (alert_status == 1):
qdb.print_out("Found valid lt surficial trigger for " + \
"%s at %s" %(site_code.upper(), ts_updated))
trigger_sym_id = qdb.get_trigger_sym_id(2, 'surficial')
df = pd.DataFrame({
'ts': [ts_updated],
'site_id': [site_id],
'trigger_sym_id': [trigger_sym_id],
'ts_updated': [ts_updated]
})
qdb.alert_to_db(df, 'operational_triggers', lt_overwrite=False)
qdb.print_out(" > Added l2 trigger on operational triggers")
# Process only l2 and l3 with alert status of -1 (invalid)
elif (alert_status == -1):
valid_cotriggers = qdb.get_valid_cotriggers(
site_id, public_ts_start)
dont_delete = False
# Check if it has co-triggers on start of event
# tho highly unlikely
if len(valid_cotriggers) != 0:
for index, valid in valid_cotriggers.iterrows():
# Don't delete public alert entry if there
# is a co-trigger that's equal or
# greater of alert level
if (valid['alert_level'] >= alert_level):
qdb.print_out(
"%s has valid co-trigger: deleting will NOT commence"
% (site_code.upper()))
dont_delete = True
break
if dont_delete == False:
qdb.delete_public_alert(site_id, public_ts_start)
qdb.print_out("Deleted {} public alert of {}".format(
public_ts_start, site_code))
def rainfall(site_id, ts, rain_id, rain_alert='both'):
"""Insert values to rainfall_alerts and operational_triggers tables
to (re)trigger rainfall alert
Args:
site_id (int): ID of site to compute rainfall analysis for.
ts (datetime): Timestamp of alert trigger.
rain_id (int): ID of rain gauge to use as data source.
rain_alert (str: {'a', 'b', 'x', None}, default None): Type of rainfall alert.
a: 1-day cumulative rainfall threshold exceeded
b: 3-day cumulative rainfall threshold exceeded
x:
None: both 1-day and 3-day cumulative rainfall threshold exceeded
"""
# 4 nearest rain gauges of each site with threshold and distance from site
gauges = rain.rainfall_gauges()
df = gauges.loc[(gauges.site_id == site_id) & (gauges.rain_id == rain_id), ['site_id', 'rain_id', 'threshold_value']]
df = df.rename(columns={'threshold_value': 'threshold'})
df.loc[:, 'ts'] = ts
# rainfall cumulative based on alert level
if str(rain_alert) == '0':
df.loc[:, 'rain_alert'] = 0
df.loc[:, 'cumulative'] = 0
else:
if rain_alert != 'x':
df.loc[:, 'rain_alert'] = 'b'
df.loc[:, 'cumulative'] = 1.2 * df.loc[:, 'threshold']
else:
df.loc[:, 'rain_alert'] = 'x'
df.loc[:, 'cumulative'] = 0.80 * df.loc[:, 'threshold']
if rain_alert == 'a' or rain_alert == 'both':
dfa = df.copy()
dfa.loc[:, ['cumulative', 'threshold']] = dfa.loc[:, ['cumulative', 'threshold']].div(2)
dfa.loc[:, 'rain_alert'] = 'a'
if rain_alert == 'a':
df = dfa.copy()
else:
df = df.append(dfa, ignore_index=True)
qdb.write_rain_alert(df)
# writes to operational_triggers
trigger_symbol = mem.get('df_trigger_symbols')
if str(rain_alert) in ['0', 'x']:
alert_level = 0
else:
alert_level = 1
trigger_sym_id = trigger_symbol.loc[(trigger_symbol.trigger_source == 'rainfall') & (trigger_symbol.alert_level == alert_level), 'trigger_sym_id'].values[0]
operational_trigger = pd.DataFrame({'site_id': [site_id], 'trigger_sym_id': [trigger_sym_id], 'ts': [ts], 'ts_updated': [ts]})
qdb.alert_to_db(operational_trigger, 'operational_triggers')
return df
def generate_surficial_alert(site_id = None,ts = None):
"""
Main alert generating function for surificial alert for a site at specified time
Parameters
------------------
site_id: int
site_id of site of interest
ts: timestamp
timestamp of alert generation
Returns
-------------------
Prints the generated alert and writes to marker_alerts database
"""
#### Obtain system arguments from command prompt
if site_id == None and ts == None:
site_id, ts = sys.argv[1].lower(),sys.argv[2].lower()
#### Config variables
num_pts = int(sc['surficial']['surficial_num_pts'])
ts_start = pd.to_datetime(ts) - timedelta(sc['surficial']['meas_plot_window'])
#### Get latest ground data
surficial_data_df = qdb.get_surficial_data(site_id, ts_start, ts, num_pts)
#### Generate Marker alerts
marker_data_df = surficial_data_df.groupby('marker_id',as_index = False)
marker_alerts = marker_data_df.apply(evaluate_marker_alerts, ts)
#### Write to marker_alerts table
data_table = sms.DataTable('marker_alerts', marker_alerts)
db.df_write(data_table)
#### Generate surficial alert for site
surficial_alert = get_surficial_alert(marker_alerts,site_id)
#### Write to db
qdb.alert_to_db(surficial_alert,'operational_triggers')
#### Plot current ground meas
if sc['surficial']['print_meas_plot']:
### Retreive the surficial data to plot
surficial_data_to_plot = surficial_data_df.loc[surficial_data_df.ts >= ts_start, :]
### Plot the surficial data
plot_site_meas(surficial_data_to_plot, ts)
return surficial_data_df
def subsurface(site_id, ts, alert_level):
"""Insert values to node_alerts, tsm_alerts, and operational_triggers
to (re)trigger subsurface alert.
Args:
site_id (int): ID of site to compute subsurface analysis for.
ts (datetime): Timestamp of alert trigger.
alert_level (int: {0, 2, 3}, default None): Subsurface alert level.
"""
# get tsm_id
query = "SELECT tsm_id FROM tsm_sensors "
query += "where site_id = {} ".format(site_id)
query += "and (date_deactivated is null or date_deactivated > '{}')".format(ts)
tsm_id = db.df_read(query, resource='sensor_data').values.flatten()
tsm_id = random.choice(tsm_id)
# writes to node_alerts; defaults to node 1 and vel_alert
ts_list = pd.date_range(end=ts, freq='30min', periods=4)
node_alerts = pd.DataFrame({'ts': ts_list, 'node_id': [1]*len(ts_list),
'tsm_id': [tsm_id]*len(ts_list),
'disp_alert': [0]*len(ts_list),
'vel_alert': [alert_level]*len(ts_list)})
db.df_write(data_table = sms.DataTable("node_alerts", node_alerts), resource='sensor_data')
# writes to tsm_alerts
tsm_alerts = pd.DataFrame({'ts': [ts], 'tsm_id': [tsm_id],
'alert_level': [alert_level],
'ts_updated': [ts]})
db.df_write(data_table = sms.DataTable("tsm_alerts", tsm_alerts), resource='sensor_data')
# writes to operational_triggers
trigger_symbol = mem.get('df_trigger_symbols')
trigger_sym_id = trigger_symbol.loc[(trigger_symbol.trigger_source == 'subsurface') & (trigger_symbol.alert_level == alert_level), 'trigger_sym_id'].values[0]
operational_trigger = pd.DataFrame({'site_id': [site_id], 'trigger_sym_id': [trigger_sym_id], 'ts': [ts], 'ts_updated': [ts]})
qdb.alert_to_db(operational_trigger, 'operational_triggers')
# details for trigger tech info
tsm_alerts.loc[:, 'node_id'] = 1
tsm_alerts.loc[:, 'disp_alert'] = 0
tsm_alerts.loc[:, 'vel_alert'] = alert_level
return tsm_alerts
def generate_surficial_alert(site_id=None,
ts=None,
marker_id=None,
to_json=False,
plot=False):
"""
Main alert generating function for surificial alert for a site at specified time
Parameters
------------------
site_id: int
site_id of site of interest
ts: timestamp
timestamp of alert generation
Returns
-------------------
Prints the generated alert and writes to marker_alerts database
"""
#### Obtain system arguments from command prompt
if site_id == None and ts == None:
site_id, ts = sys.argv[1].lower(), sys.argv[2].lower()
ts = pd.to_datetime(ts)
#### Config variables
num_pts = int(sc['surficial']['surficial_num_pts'])
ts_start = pd.to_datetime(ts) - timedelta(
sc['surficial']['meas_plot_window'])
#### Get latest ground data
surficial_data_df = qdb.get_surficial_data(site_id, ts_start, ts, num_pts)
surficial_data_df.loc[:, 'ts'] = surficial_data_df.loc[:, 'ts'].apply(
lambda x: pd.to_datetime(x))
#### Generate Marker alerts
if marker_id != None:
surficial_data_df = surficial_data_df.loc[
surficial_data_df.marker_id == marker_id, :]
marker_data_df = surficial_data_df.groupby('marker_id', as_index=False)
marker_alerts = marker_data_df.apply(evaluate_marker_alerts, ts, to_json)
#### Write to marker_alerts table
qdb.write_marker_alerts(marker_alerts[[
'ts', 'marker_id', 'data_id', 'displacement', 'time_delta',
'alert_level', 'processed'
]])
#### Generate surficial alert for site
surficial_alert = get_surficial_alert(marker_alerts, site_id)
#### Write to db
qdb.alert_to_db(surficial_alert, 'operational_triggers')
#### Plot current ground meas
if sc['surficial']['print_meas_plot'] and plot:
### Retreive the surficial data to plot
surficial_data_to_plot = surficial_data_df.loc[
surficial_data_df.ts >= ts_start, :]
### Plot the surficial data
plot_site_meas(surficial_data_to_plot, ts)
if to_json:
return marker_alerts['trend_alert'][0]
return surficial_data_df
def main(rain_props, end, sc, trigger_symbol, write_to_db=True):
"""Computes rainfall alert.
Args:
rain_props (dataframe): Contains rain gauges that can be used in
rainfall analysis.
end (datetime): Timestamp of alert to be computed.
sc (dict): Server configuration.
trigger_symbol: Alert symbol per alert level.
Returns:
dataframe: Summary of cumulative rainfall, threshold, alert and
rain gauge used in analysis of rainfall.
"""
#rainfall properties
site_id = rain_props['site_id'].values[0]
site_code = rain_props['site_code'].values[0]
twoyrmax = rain_props['threshold_value'].values[0]
halfmax = twoyrmax / 2
start = end - timedelta(float(sc['rainfall']['roll_window_length']))
offsetstart = start - timedelta(hours=0.5)
try:
if qdb.get_alert_level(site_id, end)['alert_level'].values[0] > 0:
start_monitor = pub.event_start(site_id, end)
op_trig = pub.get_operational_trigger(site_id, start_monitor, end)
op_trig = op_trig[op_trig.alert_level > 0]
validity = max(op_trig['ts_updated'].values)
validity = pub.release_time(pd.to_datetime(validity)) \
+ timedelta(1)
if 3 in op_trig['alert_level'].values:
validity += timedelta(1)
if end + timedelta(hours=0.5) >= validity:
write_alert = True
else:
write_alert = False
else:
write_alert = False
except:
write_alert = False
#data is gathered from nearest rain gauge
rainfall, gauge_name, rain_id = get_unempty_rg_data(
rain_props, offsetstart, start, end)
summary = summary_writer(site_id, site_code, gauge_name, rain_id, twoyrmax,
halfmax, rainfall, end, write_alert)
operational_trigger = summary[['site_id', 'alert']]
operational_trigger['alert'] = operational_trigger['alert'].map({
-1:
trigger_symbol[trigger_symbol.alert_level ==
-1]['trigger_sym_id'].values[0],
0:
trigger_symbol[trigger_symbol.alert_level ==
0]['trigger_sym_id'].values[0],
1:
trigger_symbol[trigger_symbol.alert_level == 1]
['trigger_sym_id'].values[0]
})
operational_trigger['ts'] = str(end)
operational_trigger['ts_updated'] = str(end)
operational_trigger = operational_trigger.rename(
columns={'alert': 'trigger_sym_id'})
if write_to_db == True:
qdb.alert_to_db(operational_trigger, 'operational_triggers')
return summary
def main(tsm_name='', end='', end_mon=False):
run_start = datetime.now()
qdb.print_out(run_start)
qdb.print_out(tsm_name)
if tsm_name == '':
tsm_name = sys.argv[1].lower()
if end == '':
try:
end = pd.to_datetime(sys.argv[2])
except:
end = datetime.now()
else:
end = pd.to_datetime(end)
window, sc = rtw.get_window(end)
tsm_props = qdb.get_tsm_list(tsm_name)[0]
data = proc.proc_data(tsm_props, window, sc)
tilt = data.tilt[window.start:window.end]
lgd = data.lgd
tilt = tilt.reset_index().sort_values('ts', ascending=True)
if lgd.empty:
qdb.print_out('%s: no data' % tsm_name)
return
nodal_tilt = tilt.groupby('node_id', as_index=False)
alert = nodal_tilt.apply(lib.node_alert,
colname=tsm_props.tsm_name,
num_nodes=tsm_props.nos,
disp=float(sc['subsurface']['disp']),
vel2=float(sc['subsurface']['vel2']),
vel3=float(sc['subsurface']['vel3']),
k_ac_ax=float(sc['subsurface']['k_ac_ax']),
lastgooddata=lgd,
window=window,
sc=sc).reset_index(drop=True)
alert.loc[:, 'col_alert'] = -1
col_alert = pd.DataFrame({
'node_id': range(1, tsm_props.nos + 1),
'col_alert': [-1] * tsm_props.nos
})
node_col_alert = col_alert.groupby('node_id', as_index=False)
node_col_alert.apply(lib.column_alert,
alert=alert,
num_nodes_to_check=int(
sc['subsurface']['num_nodes_to_check']),
k_ac_ax=float(sc['subsurface']['k_ac_ax']),
vel2=float(sc['subsurface']['vel2']),
vel3=float(sc['subsurface']['vel3']))
valid_nodes_alert = alert.loc[~alert.node_id.isin(data.inv)]
if max(valid_nodes_alert['col_alert'].values) > 0:
pos_alert = valid_nodes_alert[valid_nodes_alert.col_alert > 0]
site_alert = trend.main(pos_alert, tsm_props.tsm_id, window.end,
data.inv)
else:
site_alert = max(
lib.get_mode(list(valid_nodes_alert['col_alert'].values)))
tsm_alert = pd.DataFrame({
'ts': [window.end],
'tsm_id': [tsm_props.tsm_id],
'alert_level': [site_alert],
'ts_updated': [window.end]
})
qdb.alert_to_db(tsm_alert, 'tsm_alerts')
qdb.write_op_trig(tsm_props.site_id, window.end)
qdb.print_out(tsm_alert)
qdb.print_out('run time = ' + str(datetime.now() - run_start))
return tilt
def surficial(site_id, ts, alert_level):
"""Insert values to marker_observations, marker_data, marker_alerts, and
operational_triggers to (re)trigger surficial alert.
Args:
site_id (int): ID of site to compute surficial analysis for.
ts (datetime): Timestamp of alert trigger.
alert_level (int: {0, 1, 2, 3}, default None): Surficial alert level.
"""
# get last data for site_id
conn = mem.get('DICT_DB_CONNECTIONS')
query = "SELECT ts, marker_id, marker_name, measurement "
query += "FROM {analysis}.marker_observations "
query += "INNER JOIN {common}.sites USING (site_id) "
query += "INNER JOIN {analysis}.marker_data using (mo_id) "
query += "INNER JOIN (SELECT data_id, displacement, time_delta, alert_level, processed FROM {analysis}.marker_alerts) sub1 USING (data_id) "
query += "INNER JOIN (SELECT marker_id, marker_name FROM {analysis}.view_marker_history) sub2 USING (marker_id) "
query += "WHERE site_id = {site_id} "
query += "AND ts IN ( "
query += " SELECT MAX(ts) FROM {analysis}.marker_observations "
query += " WHERE ts < '{ts}' "
query += " AND site_id = {site_id})"
query = query.format(analysis=conn['analysis']['schema'], common=conn['common']['schema'], site_id=site_id, ts=ts)
df = db.df_read(query, resource='sensor_analysis')
# compute diff in measurements to reach threshold
if alert_level == 3:
rate = 1.8
elif alert_level in (1,2):
rate = 0.25
else:
rate = 0
meas_diff = np.ceil(rate * (ts-df.ts[0]).total_seconds()/3600)
# input measurements in inbox
gndmeas = df.loc[:, ['marker_id', 'marker_name', 'measurement']]
gndmeas.loc[:, 'ts'] = ts
gndmeas.loc[:, 'measurement'] += meas_diff
if alert_level == 1:
temp_gndmeas = gndmeas.copy()
temp_gndmeas.loc[:, 'ts'] -= (ts - df.ts[0])/2
temp_gndmeas.loc[:, 'measurement'] += meas_diff
# filler measurement for alert level 1
df_obv = pd.DataFrame({'meas_type': ['ROUTINE'], 'site_id': [site_id],
'weather': ['MAARAW'], 'observer_name':['TOPSSOFTWAREINFRA'],
'reliability': [1], 'data_source': ['SMS'],
'ts': [temp_gndmeas.ts[0]]})
mo_id = int(db.df_write(data_table=sms.DataTable("marker_observations",
df_obv), resource='sensor_data', last_insert=True)[0][0])
temp_gndmeas.loc[:, 'mo_id'] = mo_id
df_data = temp_gndmeas.loc[:, ['mo_id', 'marker_id', 'measurement']]
db.df_write(data_table = sms.DataTable("marker_data", df_data), resource='sensor_data')
surf.generate_surficial_alert(site_id = site_id, ts = temp_gndmeas.ts[0])
# measurement for ts given
df_obv = pd.DataFrame({'meas_type': ['ROUTINE'], 'site_id': [site_id],
'weather': ['MAARAW'], 'observer_name':['TOPSSOFTWAREINFRA'],
'reliability': [1], 'data_source': ['SMS'],
'ts': [ts]})
mo_id = int(db.df_write(data_table=sms.DataTable("marker_observations",
df_obv), resource='sensor_data', last_insert=True)[0][0])
gndmeas.loc[:, 'mo_id'] = mo_id
df_data = gndmeas.loc[:, ['mo_id', 'marker_id', 'measurement']]
db.df_write(data_table = sms.DataTable("marker_data", df_data), resource='sensor_data')
surf.generate_surficial_alert(site_id = site_id, ts = ts)
# details for trigger tech info
time_delta = np.round((ts - df.ts[0]).total_seconds()/3600, 2)
if alert_level == 1:
time_delta /= 2
gndmeas.loc[:, 'displacement'] = meas_diff
gndmeas.loc[:, 'time_delta'] = time_delta
gndmeas.loc[:, 'alert_level'] = alert_level
# writes to operational_triggers
trigger_symbol = mem.get('df_trigger_symbols')
trigger_sym_id = trigger_symbol.loc[(trigger_symbol.trigger_source == 'surficial') & (trigger_symbol.alert_level == alert_level), 'trigger_sym_id'].values[0]
operational_trigger = pd.DataFrame({'site_id': [site_id], 'trigger_sym_id': [trigger_sym_id], 'ts': [ts], 'ts_updated': [ts]})
qdb.alert_to_db(operational_trigger, 'operational_triggers')
return gndmeas
