This example explores how to use a BMI implementation using the Waves model as an example.
Some magic that allows us to view images within the notebook.
In [ ]:
%matplotlib inline
Import the Waves class, and instantiate it. In Python, a model with a BMI will have no arguments for its constructor. Note that although the class has been instantiated, it's not yet ready to be run. We'll get to that later!
In [ ]:
from cmt.components import Waves
waves = Waves()
Even though we can't run our waves model yet, we can still get some information about it. Just don't try to run it. Some things we can do with our model are get the names of the input variables.
In [ ]:
waves.get_output_var_names()
Or the output variables.
In [ ]:
waves.get_input_var_names()
We can also get information about specific variables. Here we'll look at some info about wave direction. This is the main output of the Waves model. Notice that BMI components always use CSDMS standard names. The CSDMS Standard Name for wave angle is,
"sea_surface_water_wave__azimuth_angle_of_opposite_of_phase_velocity"
Quite a mouthful, I know. With that name we can get information about that variable and the grid that it is on (it's actually not a one).
In [ ]:
angle_name = 'sea_surface_water_wave__azimuth_angle_of_opposite_of_phase_velocity'
print "Data type: %s" % waves.get_var_type(angle_name)
print "Units: %s" % waves.get_var_units(angle_name)
print "Grid id: %d" % waves.get_var_grid(angle_name)
print "Number of elements in grid: %d" % waves.get_grid_size(0)
print "Type of grid: %s" % waves.get_grid_type(0)
OK. We're finally ready to run the model. Well not quite. First we initialize the model with the BMI initialize method. Normally we would pass it a string that represents the name of an input file. For this example we'll pass None, which tells Waves to use some defaults.
In [ ]:
waves.initialize(None)
Before running the model, let's set a couple input parameters. These two parameters represent the frequency for which waves approach the shore at a high angle and if they come from a prefered direction.
In [ ]:
waves.set_value('sea_shoreline_wave~incoming~deepwater__ashton_et_al_approach_angle_asymmetry_parameter', .25)
waves.set_value('sea_shoreline_wave~incoming~deepwater__ashton_et_al_approach_angle_highness_parameter', .7)
To advance the model in time, we use the update method. We'll advance the model one day.
In [ ]:
waves.update()
Let's double-check that the model advanced to the given time and see what the new wave angle is.
In [ ]:
print 'Current model time: %f' % waves.get_current_time()
val = waves.get_value(angle_name)
print 'The current wave angle is: %f' % val[0]
We'll put all this in a loop and advance the model in time to generate a time series of waves angles.
In [ ]:
import numpy as np
number_of_time_steps = 400
angles = np.empty(number_of_time_steps)
for time in xrange(number_of_time_steps):
waves.update()
angles[time] = waves.get_value(angle_name)
In [ ]:
import matplotlib.pyplot as plt
plt.plot(np.array(angles) * 180 / np.pi)
plt.xlabel('Time (days)')
plt.ylabel('Incoming wave angle (degrees)')
In [ ]:
plt.hist(np.array(angles) * 180 / np.pi, bins=25)
plt.xlabel('Incoming wave angle (degrees)')
plt.ylabel('Number of occurences')