[sin123]

I am trying to write a code for orthogonal decompositions using the modred module in python. My custom vector class is defined as:

class CustomVector(mr.Vector):
def __init__(self, data_array):
self.data_array = data_array
def inner_product(self, other):
return np.dot(self.data_array, other.data_array)

Another class definition is given as:

class CustomVecHandle(mr.VecHandle):
def __init__(self, vec_path, base_handle=None, scale=None):
mr.VecHandle.__init__(self, base_handle, scale)
self.vec_path = vec_path self.res_path = result_root
def _get(self):
# read in the values
print ("reading data from {}".format(self.vec_path))
reader = vtk.vtkPolyDataReader()
reader.SetFileName(self.vec_path)
reader.Update() data = dsa.WrapDataObject(reader.GetOutput()).PointData['U']
return CustomVector(data)

And another function is defined outside all classes as:

def inner_product(v1, v2):
return v1.inner_product(v2)

I have created a list object as:

vec_handles = [CustomVecHandle(os.path.join(data_root, d, "{}_0.vtk".format(d)))for d in dirs]

This vec_handles is given as the input for a function in the modred module, and the following operations are done in the main modred code:

test_vec = vec_handles
IP_burn = self.inner_product(test_vec, test_vec)

But when I run the following code I am getting the error that:
File "/home/sindhuja/.local/lib/python2.7/site-packages/modred/vectorspace.py", line 495,
in compute_symmetric_inner_product_mat
IP_burn = self.inner_product(test_vec, test_vec)
File "podd.py", line 58,
in inner_product return v1.inner_product(v2)
AttributeError: 'list' object has no attribute 'inner_product'

I understand that the created vec_handles is a list of objects for the class CustomVecHandle. And trying to get the method using .innerproduct(test_vec) is not possible but since it's the code of the modred module, I don't think that's wrong. So how do I solve this error?

[syavash]

Originally Posted by sin123

I am trying to write a code for orthogonal decompositions using the modred module in python. My custom vector class is defined as:

class CustomVector(mr.Vector):
def __init__(self, data_array):
self.data_array = data_array
def inner_product(self, other):
return np.dot(self.data_array, other.data_array)

Another class definition is given as:

class CustomVecHandle(mr.VecHandle):
def __init__(self, vec_path, base_handle=None, scale=None):
mr.VecHandle.__init__(self, base_handle, scale)
self.vec_path = vec_path self.res_path = result_root
def _get(self):
# read in the values
print ("reading data from {}".format(self.vec_path))
reader = vtk.vtkPolyDataReader()
reader.SetFileName(self.vec_path)
reader.Update() data = dsa.WrapDataObject(reader.GetOutput()).PointData['U']
return CustomVector(data)

And another function is defined outside all classes as:

def inner_product(v1, v2):
return v1.inner_product(v2)

I have created a list object as:

vec_handles = [CustomVecHandle(os.path.join(data_root, d, "{}_0.vtk".format(d)))for d in dirs]

This vec_handles is given as the input for a function in the modred module, and the following operations are done in the main modred code:

test_vec = vec_handles
IP_burn = self.inner_product(test_vec, test_vec)

But when I run the following code I am getting the error that:
File "/home/sindhuja/.local/lib/python2.7/site-packages/modred/vectorspace.py", line 495,
in compute_symmetric_inner_product_mat
IP_burn = self.inner_product(test_vec, test_vec)
File "podd.py", line 58,
in inner_product return v1.inner_product(v2)
AttributeError: 'list' object has no attribute 'inner_product'

I understand that the created vec_handles is a list of objects for the class CustomVecHandle. And trying to get the method using .innerproduct(test_vec) is not possible but since it's the code of the modred module, I don't think that's wrong. So how do I solve this error?

Hi,

It's interesting because I was trying to do something similar and I noted the same thing in modred documentation.

Did you find the explanation as to why that line is there?
Could you please share your custom vector class?

syavash

P.S. I came to this understanding that "inner_product(v1, v2)" should be explicitly defined, as per https://modred.readthedocs.io/en/sta...ml#sec-details.
I guess it could be defined such that it return a numpy function such as numpy.dot.