def __call__(self, samples, opts=dict()):
eval_type = opts.get('eval_type', 'value')
if eval_type == 'value':
return evaluate_1darray_function_on_2d_array(
self.value, samples, opts)
elif eval_type == 'value_grad':
vals = evaluate_1darray_function_on_2d_array(
self.value, samples, opts)
return np.hstack((vals, self.gradient_set(samples).T))
elif eval_type == 'grad':
return self.gradient_set(samples).T
else:
raise Exception('%s is not a valid eval_type' % eval_type)
def value(self, samples, opts=dict()):
eval_type=opts.get('eval_type','value')
if ( ('grad' in eval_type) and
(self.func_type == "discontinuous" or self.func_type == "continuous") ):
msg = "gradients cannot be computed for %s Genz function"%self.func_type
raise Exception(msg)
assert samples.min()>=0 and samples.max()<=1.
vals = evaluate_1darray_function_on_2d_array(self.value_,samples,{})
if eval_type=='value':
return vals[:,:1]
if eval_type=='value-grad':
return vals
if eval_type=='grad':
return vals[:,1:]
def __call__(self, samples):
return evaluate_1darray_function_on_2d_array(
self.value, samples, None)
def pyapprox_fun_2(samples):
return evaluate_1darray_function_on_2d_array(fun_pause_2, samples)
def pyapprox_fun_0(samples):
values = evaluate_1darray_function_on_2d_array(fun_0, samples)
return values
def __call__(self, samples, opts):
return evaluate_1darray_function_on_2d_array(self.run, samples, opts)
