#!/usr/bin/env python
"""
Drifter
==================================
"""

from datetime import timedelta
import numpy as np
from opendrift.readers import reader_netCDF_CF_generic
from opendrift.models.oceandrift import OceanDrift

o = OceanDrift(loglevel=20)  # Basic drift model suitable for passive tracers or drifters

#%%
# Preparing Readers

reader_current = reader_netCDF_CF_generic.Reader(o.test_data_folder() +
    '16Nov2015_NorKyst_z_surface/norkyst800_subset_16Nov2015.nc')
reader_wind = reader_netCDF_CF_generic.Reader(o.test_data_folder() +
    '16Nov2015_NorKyst_z_surface/arome_subset_16Nov2015.nc')

o.add_reader([reader_current, reader_wind])

# Prevent mixing elements downwards
o.set_config('drift:vertical_mixing', False)

#%%
# Seeding elements
#
# Elements are moved with the ocean current, in addition to a fraction of
# the wind speed (wind_drift_factor). This factor depends on the properties
# of the elements. Typical empirical values are:
# - 0.035 (3.5 %) for oil and iSphere driftes
# - 0.01  (1 %) for CODE drifters partly submerged ~0.5 m
# As there are large uncertainties, it makes sense to provide a statistical
# distribution of wind_drift_factors
#
# Using a constant value for all elements:
#wind_drift_factor = 0.03
#
# Giving each element a unique (random) wind_drift_factor
wind_drift_factor = np.random.uniform(0, 0.06, 2000)
o.seed_elements(4.7, 59.9, radius=3000, number=2000,
                time=reader_current.start_time,
                wind_drift_factor=wind_drift_factor)

#%%
# Running model
o.run(time_step=timedelta(minutes=15),
      time_step_output=timedelta(minutes=60))

#%%
# Print and plot results
print(o)
o.animation(color='wind_drift_factor', fast=True)

#%%
# .. image:: /gallery/animations/example_drifter_0.gif

# Plot trajectories, colored by the wind_drift_factor of each element
o.plot(linecolor='wind_drift_factor', fast=True)