Lanczos_ER doesn't work for type f, cf, cd vector

[j9263178]

An example in python (cd type LinOp and input vector):

class Htest(cytnx.LinOp):

    def __init__(self, psidim):
        cytnx.LinOp.__init__(self,"mv", psidim, cytnx.Type.ComplexDouble, cytnx.Device.cpu)

    def matvec(self, v):
        v_ = v.clone()
        return 1.02*v_

if __name__ == '__main__':
  vec = cytnx.arange(1,5,1, cytnx.Type.ComplexDouble, cytnx.Device.cpu)
  _,_ = cytnx.linalg.Lanczos_ER(Htest(len(vec)), maxiter = 4, CvgCrit = 9999999999, Tin = vec, max_krydim = krydim)

Result:

.....

File "/home/j9263178/cytnx_test/cytnx/linalg_conti.py", line 7, in Lanczos_ER
    return linalg.c_Lanczos_ER(Hop,k,is_V,maxiter,CvgCrit,is_row,Tin,max_krydim,verbose);
RuntimeError: 
# Cytnx error occur at void cytnx::Tensor_impl::Init(const std::vector<long unsigned int>&, const unsigned int&, int)
# error: [ERROR] invalid argument: dtype
# file : /home/j9263178/github/Cytnx/src/Tensor.cpp (314)

Result for cf type:

程式記憶體區段錯誤 (core dumped)

Result for f type:

.....

 File "/home/j9263178/cytnx_test/cytnx/linalg_conti.py", line 7, in Lanczos_ER
    return linalg.c_Lanczos_ER(Hop,k,is_V,maxiter,CvgCrit,is_row,Tin,max_krydim,verbose);
RuntimeError: 
# Cytnx error occur at T& cytnx::Storage_base::at(const cytnx_uint64&) const [with T = float; cytnx::cytnx_uint64 = long unsigned int]
# error: [ERROR] type mismatch. try to get <float> type from raw data of type Double (Float64)
# file : /home/j9263178/github/Cytnx/src/Storage_base.cpp (625)

[kaihsin]

please refer the full fix on 76cf976