def period_op(ts, period="D", agg="mean", max_absent_frac=0.):
period = pd.tseries.frequencies.to_offset(period)
# resample getting a sum and non-null count
ts1 = ts.resample(period).agg([agg, 'count']).copy()
# determine required number data which is exact for period of "D" or smaller
agged = ts1.loc[:, (slice(None), agg)].droplevel(level=1, axis=1)
if max_absent_frac is None:
return agged
counts = ts1.loc[:, (slice(None), 'count')].droplevel(level=1, axis=1)
try:
na_tol = (period / ts.index.freq) * (1. - max_absent_frac)
invalid = counts < na_tol
except:
if period == months(1):
# determine invalid months
na_tol = max_absent_frac * ts.index.days_in_month * (days(1) /
ts.index.freq)
invalid = counts < na_tol
else:
raise ValueError(
"Offset {} not supported. Only absolute offsets (<= 1day) or one month averages supported with max_absent_frac option"
.format(period))
agged.mask(invalid, inplace=True)
return agged
def plot_inst_and_avg(axes, tss, window_inst, window_avg, labels, label_loc, legend_size):
""" Plot instantaneous and filtered time series plot
"""
if window_inst is None:
window_inst = get_union_window(tss)
lines = plot_tss(axes['inst'], tss, window_inst)
if labels is not None:
axes['inst'].legend(lines, labels, prop={'size': legend_size}, loc=label_loc)
if window_avg is None:
window_avg = get_union_window(tss)
pad = days(4)
window_to_filter = (window_avg[0] - pad, window_avg[1] + pad)
tss_clipped = [safe_window(ts, window_to_filter) for ts in tss]
tss_filtered = filter_timeseries(tss_clipped)
plot_tss(axes['avg'], tss_filtered, window_avg)
def main():
parser = create_arg_parser()
args = parser.parse_args()
monterey_fpath = args.monterey
pt_reyes_fpath = args.pt_reyes
hgrid_fpath = args.hgrid
fpath_out = args.outfile
#todo: hardwire
nnode = 83
tbuf = days(16)
# convert start time string input to datetime
sdate = datetime(*map(int, re.split('[^\d]', args.stime)))
if args.etime:
# convert start time string input to datetime
edate = datetime(*map(int, re.split('[^\d]', args.etime)))
bufend = edate + tbuf
else:
edate = None
bufend = None
# UTM positions of Point Reyes, Monterey, SF
pos_pr = np.array([502195.03, 4205445.47])
pos_mt = np.array([599422.84, 4051630.37])
pos_sf = np.array([547094.79, 4184499.42])
var_subtidal = np.array([0.938, 0.905, 0.969]) # pr, mt, sf
var_semi = np.array([0.554, 0.493, 0.580])
# Assume 45 degree from north-west to south-east
tangent = np.array([1, -1])
tangent = tangent / np.linalg.norm(tangent) # Normalize
# Rotate 90 cw to get normal vec
normal = np.array([tangent[1], -tangent[0]])
print "tangent:", tangent
print "normal:", normal
mt_rel = pos_mt - pos_pr
x_mt = np.dot(tangent, mt_rel) # In pr-mt direction
y_mt = np.dot(normal, mt_rel) # Normal to x-direction to the
# Grid
#todo: what is the difference between this and m = read_grid()??
s = create_schism_setup(hgrid_fpath)
ocean_boundary = s.mesh.boundaries[0] # First one is ocean
# Data
print "Reading Point Reyes..."
pt_reyes = read_ts(pt_reyes_fpath, start=sdate - tbuf, end=bufend, force_regular=True)
pt_reyes = interpolate_ts_nan(pt_reyes)
if np.any(np.isnan(pt_reyes.data)):
print pt_reyes.times[np.isnan(pt_reyes.data)]
ts_pr_subtidal, ts_pr_diurnal, ts_pr_semi, noise = separate_species(
pt_reyes)
del noise
print "Reading Monterey..."
monterey = read_ts(monterey_fpath, start=sdate - tbuf, end=bufend, force_regular=True)
monterey = interpolate_ts_nan(monterey)
if pt_reyes.interval != monterey.interval:
raise ValueError(
"Point Reyes and Monterey time step must be the same in gen_elev2D.py")
ts_mt_subtidal, ts_mt_diurnal, ts_mt_semi, noise = separate_species(
monterey)
del noise
dt = monterey.interval.seconds
print "Done Reading"
print "Interpolating Point Reyes"
# interpolate_ts(ts_pr_subtidal.window(sdate,edate),step)
ts_pr_subtidal = ts_pr_subtidal.window(sdate, edate)
ts_pr_diurnal = ts_pr_diurnal.window(sdate, edate) # interpolate_ts(,step)
# interpolate_ts(ts_pr_semi.window(sdate,edate),step)
ts_pr_semi = ts_pr_semi.window(sdate, edate)
print "Interpolating Monterey"
# interpolate_ts(ts_mt_subtidal.window(sdate,edate),step)
ts_mt_subtidal = ts_mt_subtidal.window(sdate, edate)
# interpolate_ts(ts_mt_diurnal.window(sdate,edate),step)
ts_mt_diurnal = ts_mt_diurnal.window(sdate, edate)
# interpolate_ts(ts_mt_semi.window(sdate,edate),step)
ts_mt_semi = ts_mt_semi.window(sdate, edate)
print "Creating writer" # requires dt be known for netcdf
if fpath_out.endswith("th"):
thwriter = BinaryTHWriter(fpath_out,nnode,None)
elif fpath_out.endswith("nc"):
thwriter = NetCDFTHWriter(fpath_out,nnode,sdate,dt)
else:
raise ValueError("File extension for output not recognized in file: {}".format(fpath_out))
# Grid
boundaries = s.mesh.nodes[ocean_boundary.nodes]
pos_rel = boundaries[:, :2] - pos_pr
# x, y in a new principal axes
x = np.dot(pos_rel, tangent.reshape((2, -1)))
y = np.dot(pos_rel, normal.reshape((2, -1)))
theta_x = x / x_mt
theta_x_comp = 1. - theta_x
theta_y = y / y_mt
theta_y_comp = 1. - theta_y
var_y = (theta_y_comp * var_semi[0] + theta_y * var_semi[1])
# adj_subtidal_mt = 0.08 # Adjustment in Monterey subtidal signal
# scaling_diurnal_mt = 0.95 # Scaling of Monterey diurnal signal (for K1/Q1)
# Used this up to v75
adj_subtidal_mt = 0. # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 1. # Scaling of Monterey diurnal signal (for K1/Q1)
# New trial for LSC2 with v75
adj_subtidal_mt = -0.07 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.95 # Scaling of Monterey diurnal signal (for K1/Q1)
scaling_semidiurnal_mt = 1.03
adj_subtidal_mt = -0.14 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.90 # Scaling of Monterey diurnal signal (for K1/Q1)
scaling_semidiurnal_mt = 1.07
adj_subtidal_mt = 0.10 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.90 # Scaling of Monterey diurnal signal (for K1/Q1)
scaling_semidiurnal_mt = 1.03
adj_subtidal_mt = 0.10 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.97 # Scaling of Monterey diurnal signal (for K1/Q1)
# Scaling of Point Reyes diurnal signal (for K1/Q1)
scaling_diurnal_pr = 0.97
scaling_semidiurnal_mt = 1.025 # Scaling at Monterey semi-diurnal signal
adj_subtidal_mt = 0.09 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.94 # Scaling of Monterey diurnal signal (for K1/Q1)
# Scaling of Point Reyes diurnal signal (for K1/Q1)
scaling_diurnal_pr = 0.94
scaling_semidiurnal_mt = 1.0 # Scaling at Monterey semi-diurnal signal
slr = 0.0 # Sea level rise
if np.any(np.isnan(ts_pr_semi.data)):
print ts_pr_semi.times[np.isnan(ts_pr_semi.data)]
raise ValueError('')
# temp=np.zeros((len(ts_pr_semi),nnode))
# times[:]=[dt*i for i in xrange(len(ts_pr_semi))]
for i in xrange(len(ts_pr_semi)):
t = float(dt * i)
# semi-diurnal
# Scaling
pr = ts_pr_semi[i].value
mt = ts_mt_semi[i].value * scaling_semidiurnal_mt
if np.isnan(pr) or np.isnan(mt):
print pr
print mt
raise ValueError("One of values is numpy.nan.")
eta_pr_side = var_y / var_semi[0] * pr
eta_mt_side = var_y / var_semi[1] * mt
eta = eta_pr_side * theta_x_comp + eta_mt_side * theta_x
# diurnal
# Interpolate in x-direction only to get a better phase
pr = ts_pr_diurnal[i].value * scaling_diurnal_pr
mt = ts_mt_diurnal[i].value * scaling_diurnal_mt
if np.isnan(pr) or np.isnan(mt):
raise ValueError("One of values is numpy.nan.")
eta += pr * theta_x_comp + mt * theta_x
# Subtidal
# No phase change in x-direction. Simply interpolate in
# y-direction.
pr = ts_pr_subtidal[i].value + slr
mt = ts_mt_subtidal[i].value + adj_subtidal_mt + slr
if np.isnan(pr) or np.isnan(mt):
raise ValueError("One of values is numpy.nan.")
eta += pr * theta_y_comp + mt * theta_y
# write data to netCDF file
thwriter.write_step(i,t,eta)
# Delete class
del thwriter
def read(self, fpath, start=None, end=None, force_regular=True, selector=None):
""" Read a text file with the given pattern and parsers.
Parsers and a pattern must be defined and set in the child class.
Parameters
----------
fpath: str
file to read
start: datetime.datetime, optional
datetime to start reading in.
If None, read from the start of the file
end: datetime.datetime, optional
datetime to finish reading in.
If None, read till the end of the file
force_regular: boolean, optional
If it is true, it returns a regular time series
Returns
-------
vtools.data.timeseries.TimeSeries
time series from the file
"""
# The selector (if it exists) can probably be precalculated or at least recorded.
# Almost always this amounts to picking variables out of a list of column names
# and recording indexes, but here we don't ask any questions about what "selector" is.
n_headerlines, metadata = self.process_header(fpath,selector)
if n_headerlines is None: raise ValueError("Problem counting header lines (check format?)")
metadata = dict()
if not self._header_regexs is None:
metadata = self.read_metadata_from_header(fpath)
if not start is None and not end is None:
if start >= end:
raise ValueError("The end time must be later than the start")
with open(fpath, 'r') as f_in:
times = list()
values = list()
# fast forward past header
if n_headerlines > 0:
for _ in range(n_headerlines):
f_in.readline()
# process lines starting from current file pointer
for i, line in enumerate(f_in):
if self.is_comment(line):
continue
timestamp, vals = self.parse_record(line)
if start and timestamp < start:
continue
if end and timestamp > end:
break
times.append(timestamp)
values.append(vals)
if len(times) < 1:
return None
arr = numpy.array(values)
# Here I assumed that it is more effective to retrieve too much
# in the reading stage and then do this with numpy fancy indexing.
# I But you can override this function
# todo: never seemed complete selector handled elsewhere
#arr = self.cull_using_selector(arr)
ts = vts.its(times, arr)
if force_regular:
interval = infer_interval(times[:11],
fraction=0.5,
standard=[vtt.minutes(6),
vtt.minutes(10),
vtt.minutes(15),
vtt.hours(1),
vtt.days(1)])
if not interval:
# for t in times[:10]:
# print t.strftime("%Y-%m-%d %H:%M:%S")
raise ValueError("Interval could not be inferred from first time steps in %s" % fpath)
import warnings
# todo: this really should be an option
with warnings.catch_warnings():
warnings.simplefilter("ignore")
ts = vts.its2rts(ts, interval)
# if start is not None:
# if start < ts.start:
# print "extending front..."
# ts = vts.extrapolate_ts(ts, start=start)
# print ts.times, ts.data
# else:
# ts = ts.window(start=start)
# if end is not None:
# if end > ts.end:
# print "extending back..."
# ts = vts.extrapolate_ts(ts, end=end)
# else:
# ts = ts.window(end=end)
for k, v in metadata.items():
ts.props[k] = v
return ts
def plot(self):
""" Generate metrics plots
"""
# Process input parameters
params = self.params
variable = params['variable']
outputs_dir = params['outputs_dir']
if isinstance(outputs_dir, str):
outputs_dir = outputs_dir.split()
time_basis = process_time_str(params['time_basis'])
stations_input = params.get('stations_input')
if stations_input is None:
stations_input = params.get(
'flow_station_input') if variable == "flow" else params.get(
'station_input')
else:
raise ValueError(
"Old style input file. \nUse 'station_input' and 'flow_station_input' respectively for staout* and flow.dat"
)
if isinstance(stations_input, str):
stations_input = stations_input.split()
db_stations = station.read_station_dbase(params['stations_csv'])
db_obs = station.read_obs_links(params['obs_links_csv'])
excluded_stations = params.get('excluded_stations')
selected_stations = params.get('selected_stations')
start_avg = process_time_str(params["start_avg"])
end_avg = process_time_str(params["end_avg"])
start_inst = process_time_str(params["start_inst"])
end_inst = process_time_str(params["end_inst"])
labels = params['labels']
dest_dir = params.get('dest_dir')
if dest_dir is None:
dest_dir = '.'
else:
if not os.path.exists(dest_dir):
os.mkdir(dest_dir)
plot_format = params.get('plot_format')
padding = days(4)
window_common = (min(start_inst, start_avg), max(end_inst, end_avg))
window_to_read = (window_common[0] - padding,
window_common[1] + padding)
plot_all = read_optional_flag_param(params, 'plot_all')
remove_outliers = read_optional_flag_param(params, 'remove_outliers')
adjust_datum = read_optional_flag_param(params, 'auto_adjustment')
fill_gap = read_optional_flag_param(params, 'fill_gap')
max_gap_to_fill = hours(1)
if 'max_gap_to_fill' in params:
max_gap_to_fill = pd.tseries.frequencies.to_offset(
params['max_gap_to_fill'])
else:
max_gap_to_fill = hours(1)
# Prepare readers of simulation outputs
sim_outputs = self.read_simulation_outputs(variable, outputs_dir,
time_basis, stations_input)
assert len(sim_outputs) > 0
assert sim_outputs[0] is not None
# Iterate through the stations in the first simulation outputs
for stn in sim_outputs[0].columns:
station_id = stn[0] if type(stn) == tuple else stn
# Prepare
self.logger.info(
"==================================================")
self.logger.info(
"Start processing station:: {}".format(station_id))
if not station_id in db_stations.index:
self.logger.warning(
"Station id {} not found in station listings".format(
station_id))
continue
alias = db_stations.loc[station_id, 'alias']
if selected_stations is not None:
if station_id not in selected_stations:
self.logger.info(
"Skipping..."
" Not in the list of the selected stations: %s",
station_id)
continue
if excluded_stations is not None:
if station_id in excluded_stations:
self.logger.info(
"Skipping... "
"In the list of the excluded stations: %s", station_id)
continue
if variable == 'flow':
vert_pos = 'default'
else:
vert_pos = stn[1]
adj_obs = 0.
# Read Obs
subloc = 'default' if variable == 'flow' else stn[1]
ts_obs = self.retrieve_ts_obs(station_id, subloc, variable,
window_to_read, db_stations, db_obs)
if ts_obs is None or ts_obs.isnull().all():
self.logger.warning("No observation data: %s.", station_id)
if plot_all is False:
self.logger.warning("Skipping this station")
continue
else:
if remove_outliers is True:
self.logger.info("Removing outliers...")
ts_obs = med_outliers(ts_obs, level=3, copy=False)
adj = db_obs.loc[(station_id, subloc, variable), 'datum_adj']
if adj is not None and adj != 0.:
self.logger.info(
"Adjusting obs value with the value in the table...")
ts_obs += adj
if obs_unit == 'ft':
adj = ft_to_m(adj)
else:
ValueError("Not supported unit for adjustment.")
adj_obs += adj
try:
obs_unit = db_obs.loc[(station_id, subloc, variable),
'unit']
ts_obs = self.convert_unit_of_ts_obs_to_SI(
ts_obs, obs_unit)
obs_unit = ts_obs.unit
except Exception as e:
raise Exception("Station {}".format(station_id)) from e
# Read simulation
if variable == "flow":
tss_sim = [
None if simout[station_id].isnull().all() else
simout[station_id] for simout in sim_outputs
]
else:
tss_sim = [
simout.loc[:, (station_id, subloc)].iloc[:, 0]
for simout in sim_outputs
]
for ts in tss_sim:
if ts is None: continue
if ts_obs is None or ts_obs.isnull().all():
ts.unit = self.variable_units[variable]
else:
ts.unit = obs_unit
# Adjust datum if necessary
if adjust_datum and ts_obs is not None:
ts_obs, adj = self.adjust_obs_datum(ts_obs, tss_sim[0],
station_id, variable,
db_obs)
adj_obs += adj
if ts_obs is not None and fill_gap is True:
self.logger.info("Filling gaps in the data.")
ts_obs = fill_gaps(ts_obs, max_gap_to_fill)
# Plot
if check_if_all_tss_are_bad([ts_obs] + tss_sim):
self.logger.error("None of time series have data.")
continue
self.logger.info("Start plotting...")
source = db_obs.loc[(station_id, subloc, variable),
"agency"].upper()
figtitle = self.create_title(db_stations, station_id, source,
variable, vert_pos)
title = None
if type(tss_sim) == list: tss_sim = tuple(tss_sim)
# labels
labels_to_plot = deepcopy(labels)
if adj_obs != 0.:
if adj_obs > 0.:
labels_to_plot[0] += " + {:g}".format(adj_obs)
else:
labels_to_plot[0] += " - {:g}".format(-adj_obs)
if plot_format == 'simple':
fig = plot_comparison(ts_obs,
tss_sim,
window_inst=(start_inst, end_inst),
window_avg=(start_avg, end_avg),
labels=labels_to_plot,
title=title)
else:
fig = plot_metrics(ts_obs,
tss_sim,
window_inst=(start_inst, end_inst),
window_avg=(start_avg, end_avg),
labels=labels_to_plot,
title=title)
fname_output = self.set_fname_out(alias, variable, station_id,
vert_pos)
fpath_output = os.path.join(dest_dir, fname_output + '.png')
fig.suptitle(figtitle, fontsize=14)
fig.savefig(fpath_output, dpi=300)
self.logger.info("Done for the station.")
def plot(self):
""" Generate metrics plots
"""
# Process input parameters
params = self.params
variable = params['variable']
outputs_dir = params['outputs_dir']
if isinstance(outputs_dir, str):
outputs_dir = outputs_dir.split()
time_basis = process_time_str(params['time_basis'])
stations_input = params.get('stations_input')
if stations_input is None:
stations_input = params.get(
'flow_station_input') if variable == "flow" else params.get(
'station_input')
else:
raise ValueError(
"Old style input file. \nUse 'station_input' and 'flow_station_input' respectively for staout* and flow.dat"
)
if isinstance(stations_input, str):
stations_input = stations_input.split()
db_stations = station_db.StationDB(params['stations_csv'])
db_obs = obs_links.ObsLinks(params['obs_links_csv'])
excluded_stations = params.get('excluded_stations')
selected_stations = params.get('selected_stations')
start_avg = process_time_str(params["start_avg"])
end_avg = process_time_str(params["end_avg"])
start_inst = process_time_str(params["start_inst"])
end_inst = process_time_str(params["end_inst"])
labels = params['labels']
dest_dir = params.get('dest_dir')
if dest_dir is None:
dest_dir = '.'
else:
if not os.path.exists(dest_dir):
os.mkdir(dest_dir)
plot_format = params.get('plot_format')
padding = days(4)
window_common = (min(start_inst, start_avg), max(end_inst, end_avg))
window_to_read = (window_common[0] - padding,
window_common[1] + padding)
plot_all = read_optional_flag_param(params, 'plot_all')
remove_outliers = read_optional_flag_param(params, 'remove_outliers')
adjust_datum = read_optional_flag_param(params, 'auto_adjustment')
fill_gap = read_optional_flag_param(params, 'fill_gap')
max_gap_to_fill = hours(1)
if 'max_gap_to_fill' in params:
max_gap_to_fill = parse_interval(params['max_gap_to_fill'])
# Prepare readers of simulation outputs
sim_outputs = self.read_simulation_outputs(variable, outputs_dir,
time_basis, stations_input)
assert len(sim_outputs) > 0
assert sim_outputs[0] is not None
# Iterate through the stations in the first simulation outputs
for station in sim_outputs[0].stations:
# Prepare
self.logger.info(
"==================================================")
self.logger.info("Start processing a station: %s", station["name"])
station_id = station['name']
alias = db_stations.alias(station_id)
if selected_stations is not None:
if station_id not in selected_stations:
self.logger.info(
"Skipping..."
" Not in the list of the selected stations: %s",
station_id)
continue
if excluded_stations is not None:
if station_id in excluded_stations:
self.logger.info(
"Skipping... "
"In the list of the excluded stations: %s", station_id)
continue
if not variable == 'flow':
vert_pos = station['vert_pos']
else:
vert_pos = 0
adj_obs = 0.
# Read Obs
ts_obs = self.retrieve_ts_obs(station_id, variable, window_to_read,
db_stations, db_obs, vert_pos)
if ts_obs is None:
self.logger.warning("No observation data: %s.", station_id)
if plot_all is False:
self.logger.warning("Skipping this station")
continue
else:
if remove_outliers is True:
self.logger.info("Removing outliers...")
ts_obs, filtered = med_outliers(ts_obs, copy=False)
adj = db_obs.adjustment(station_id, variable)
if adj is not None and adj != 0.:
self.logger.info(
"Adjusting obs value with the value in the table...")
ts_obs += adj
obs_unit = db_obs.unit(station_id, variable, vert_pos)
if obs_unit == 'ft':
adj = ft_to_m(adj)
else:
ValueError("Not supported unit for adjustment.")
adj_obs += adj
if 'unit' not in ts_obs.props:
ts_obs.props['unit'] = db_obs.unit(station_id, variable)
ts_obs = self.convert_unit_of_ts_obs_to_SI(ts_obs)
# Read simulation
tss_sim = self.retrieve_tss_sim(sim_outputs, station_id, variable,
vert_pos)
# Adjust datum if necessary
if adjust_datum is True and ts_obs is not None:
ts_obs, adj = self.adjust_obs_datum(ts_obs, tss_sim[0],
station_id, variable,
db_obs)
adj_obs += adj
if ts_obs is not None and fill_gap is True:
self.logger.info("Filling gaps in the data.")
fill_gaps((ts_obs, ), max_gap_to_fill)
# Plot
if check_if_all_tss_are_bad([
ts_obs,
] + tss_sim):
self.logger.error("None of time series have data.")
continue
self.logger.info("Start plotting...")
source = db_obs.agency(station_id, variable).upper()
figtitle = self.create_title(db_stations, station_id, source,
variable, vert_pos)
title = None
# labels
labels_to_plot = deepcopy(labels)
if adj_obs != 0.:
if adj_obs > 0.:
labels_to_plot[0] += " + %g" % adj_obs
else:
labels_to_plot[0] += " - %g" % (-adj_obs)
if plot_format == 'simple':
fig = plot_comparison(ts_obs,
*tss_sim,
window_inst=(start_inst, end_inst),
window_avg=(start_avg, end_avg),
labels=labels_to_plot,
title=title)
else:
fig = plot_metrics(ts_obs,
*tss_sim,
window_inst=(start_inst, end_inst),
window_avg=(start_avg, end_avg),
labels=labels_to_plot,
title=title)
fname_output = self.set_fname_out(alias, variable, station_id,
vert_pos)
fpath_output = os.path.join(dest_dir, fname_output + '.png')
fig.suptitle(figtitle, fontsize=14)
fig.savefig(fpath_output, dpi=300)
self.logger.info("Done for the station.")
# -*- coding: utf-8 -*-
from vtools.functions.period_op import *
from vtools.data.vtime import days
from vtools.data.sample_series import *
import matplotlib.pyplot as plt
ts_week=synthetic_tide_series()
## get 2 days of data
ts=ts_week.window(ts_week.times[0],ts_week.times[0]+days(2))
ts_max=period_max(ts,hours(1))
fig=plt.figure()
ax0 = fig.add_subplot(111)
ax0.set_ylabel("surface (feet)")
p0=ax0.plot(ts.times,ts.data,color='g',linewidth=1.0)
p1=ax0.step(ts_max.times,ts_max.data,color='r',linewidth=1.0,where="post")
p2=ax0.plot(ts_max.centered().times,ts_max.centered().data,color='black',\
linewidth=1.0,linestyle="--")
plt.legend(["Surface","Max_step","Max_centered"])
plt.grid(b=True, which='major', color='0.9', linestyle='-', linewidth=0.5)
fig.autofmt_xdate()
plt.show()
def gen_elev2D(hgrid_fpath, outfile, pt_reyes_fpath, monterey_fpath, start,
end, slr):
max_gap = 5
stime = start
etime = end
fpath_out = outfile
#todo: hardwire
nnode = 83
tbuf = days(16)
# convert start time string input to datetime
sdate = pd.Timestamp(stime)
if not etime is None:
# convert start time string input to datetime
edate = pd.Timestamp(etime)
bufend = edate + tbuf
else:
edate = None
bufend = None
# UTM positions of Point Reyes, Monterey, SF
pos_pr = np.array([502195.03, 4205445.47])
pos_mt = np.array([599422.84, 4051630.37])
pos_sf = np.array([547094.79, 4184499.42])
var_subtidal = np.array([0.938, 0.905, 0.969]) # pr, mt, sf
var_semi = np.array([0.554, 0.493, 0.580])
# Assume 45 degree from north-west to south-east
tangent = np.array([1, -1])
tangent = tangent / np.linalg.norm(tangent) # Normalize
# Rotate 90 cw to get normal vec
normal = np.array([tangent[1], -tangent[0]])
print("tangent: {}".format(tangent))
print("normal: {}".format(normal))
mt_rel = pos_mt - pos_pr
x_mt = np.dot(tangent, mt_rel) # In pr-mt direction
y_mt = np.dot(normal, mt_rel) # Normal to x-direction to the
# Grid
#todo: what is the difference between this and m = read_grid()??
mesh = read_mesh(hgrid_fpath)
ocean_boundary = mesh.boundaries[0] # First one is ocean
# Data
print("Reading Point Reyes...")
pt_reyes = read_noaa(pt_reyes_fpath,
start=sdate - tbuf,
end=bufend,
force_regular=True)
pt_reyes.interpolate(limit=max_gap, inplace=True)
if pt_reyes.isna().any(axis=None):
raise ValueError("pt_reyes has gaps larger than fill limit")
ts_pr_subtidal, ts_pr_diurnal, ts_pr_semi, noise = separate_species(
pt_reyes, noise_thresh_min=150)
del noise
print("Reading Monterey...")
monterey = read_noaa(monterey_fpath,
start=sdate - tbuf,
end=bufend,
force_regular=True)
monterey.interpolate(limit=max_gap, inplace=True)
if pt_reyes.isna().any(axis=None):
raise ValueError("monterey has gaps larger than fill limit")
if pt_reyes.index.freq != monterey.index.freq:
raise ValueError(
"Point Reyes and Monterey time step must be the same in gen_elev2D.py"
)
ts_mt_subtidal, ts_mt_diurnal, ts_mt_semi, noise = separate_species(
monterey, noise_thresh_min=150)
del noise
dt = monterey.index.freq / seconds(1)
print("Done Reading")
print("Interpolating and subsetting Point Reyes")
# interpolate_ts(ts_pr_subtidal.window(sdate,edate),step)
ts_pr_subtidal = ts_pr_subtidal.loc[sdate:edate]
ts_pr_diurnal = ts_pr_diurnal.loc[sdate:edate]
# interpolate_ts(ts_pr_semi.window(sdate,edate),step)
ts_pr_semi = ts_pr_semi.loc[sdate:edate]
print("Interpolating and subsetting Monterey")
# interpolate_ts(ts_mt_subtidal.window(sdate,edate),step)
ts_mt_subtidal = ts_mt_subtidal.loc[sdate:edate]
# interpolate_ts(ts_mt_diurnal.window(sdate,edate),step)
ts_mt_diurnal = ts_mt_diurnal.loc[sdate:edate]
# interpolate_ts(ts_mt_semi.window(sdate,edate),step)
ts_mt_semi = ts_mt_semi.loc[sdate:edate]
print("Creating writer") # requires dt be known for netcdf
if fpath_out.endswith("th"):
thwriter = BinaryTHWriter(fpath_out, nnode, None)
elif fpath_out.endswith("nc"):
thwriter = NetCDFTHWriter(fpath_out, nnode, sdate, dt)
else:
raise ValueError(
"File extension for output not recognized in file: {}".format(
fpath_out))
# Grid
boundaries = mesh.nodes[ocean_boundary.nodes]
pos_rel = boundaries[:, :2] - pos_pr
# x, y in a new principal axes
x = np.dot(pos_rel, tangent.reshape((2, -1)))
y = np.dot(pos_rel, normal.reshape((2, -1)))
theta_x = x / x_mt
theta_x_comp = 1. - theta_x
theta_y = y / y_mt
theta_y_comp = 1. - theta_y
var_y = (theta_y_comp * var_semi[0] + theta_y * var_semi[1])
# adj_subtidal_mt = 0.08 # Adjustment in Monterey subtidal signal
# scaling_diurnal_mt = 0.95 # Scaling of Monterey diurnal signal (for K1/Q1)
# Used this up to v75
adj_subtidal_mt = 0. # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 1. # Scaling of Monterey diurnal signal (for K1/Q1)
# New trial for LSC2 with v75
adj_subtidal_mt = -0.07 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.95 # Scaling of Monterey diurnal signal (for K1/Q1)
scaling_semidiurnal_mt = 1.03
adj_subtidal_mt = -0.14 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.90 # Scaling of Monterey diurnal signal (for K1/Q1)
scaling_semidiurnal_mt = 1.07
adj_subtidal_mt = 0.10 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.90 # Scaling of Monterey diurnal signal (for K1/Q1)
scaling_semidiurnal_mt = 1.03
adj_subtidal_mt = 0.10 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.97 # Scaling of Monterey diurnal signal (for K1/Q1)
# Scaling of Point Reyes diurnal signal (for K1/Q1)
scaling_diurnal_pr = 0.97
scaling_semidiurnal_mt = 1.025 # Scaling at Monterey semi-diurnal signal
adj_subtidal_mt = 0.09 # Adjustment in Monterey subtidal signal
scaling_diurnal_mt = 0.94 # Scaling of Monterey diurnal signal (for K1/Q1)
# Scaling of Point Reyes diurnal signal (for K1/Q1)
scaling_diurnal_pr = 0.94
scaling_semidiurnal_mt = 1.0 # Scaling at Monterey semi-diurnal signal
if ts_pr_semi.isna().any(axis=None):
print(ts_pr_semi[ts_pr_semi.isna()])
raise ValueError('Above times are missing in Point Reyes data')
for i in range(len(ts_pr_semi)):
t = float(dt * i)
# semi-diurnal
# Scaling
pr = ts_pr_semi.iloc[i, 0]
mt = ts_mt_semi.iloc[i, 0] * scaling_semidiurnal_mt
if np.isnan(pr) or np.isnan(mt):
raise ValueError("One of values is numpy.nan.")
eta_pr_side = var_y / var_semi[0] * pr
eta_mt_side = var_y / var_semi[1] * mt
eta = eta_pr_side * theta_x_comp + eta_mt_side * theta_x
# diurnal
# Interpolate in x-direction only to get a better phase
pr = ts_pr_diurnal.iloc[i, 0] * scaling_diurnal_pr
mt = ts_mt_diurnal.iloc[i, 0] * scaling_diurnal_mt
#if i < 5:
# print("yu")
# print(pr)
# print(mt)
if np.isnan(pr) or np.isnan(mt):
raise ValueError("One of values is numpy.nan.")
eta += pr * theta_x_comp + mt * theta_x
# Subtidal
# No phase change in x-direction. Simply interpolate in
# y-direction.
pr = ts_pr_subtidal.iloc[i, 0]
mt = ts_mt_subtidal.iloc[i, 0] + adj_subtidal_mt
if np.isnan(pr) or np.isnan(mt):
raise ValueError("One of values is numpy.nan.")
eta += pr * theta_y_comp + mt * theta_y + slr
# write data to netCDF file
thwriter.write_step(i, t, eta)
# Delete class
del thwriter
