def __init__(self, d, ell):
self.class_name = 'EntrySampler'
self.d = d
self.ell = ell
self.nnz = d*ell
self.rows = 0
self.sampler = ReservoirSampler(self.nnz)
class EntrySampler:
def __init__(self, d, ell):
self.class_name = 'EntrySampler'
self.d = d
self.ell = ell
self.nnz = d*ell
self.rows = 0
self.sampler = ReservoirSampler(self.nnz)
def append(self,v):
for (col,val) in enumerate(v):
self.sampler.add((self.rows, col, val), abs(val))
self.rows += 1
def get(self):
B = dok_matrix((self.rows,self.d), dtype=float32)
for ((row,col,val),p) in self.sampler.get(with_probabilities=True):
B[row,col] += val/(p*self.nnz)
covariance = dot(B.transpose(),B)
(_,s,Vt) = svds(covariance, k=self.ell, maxiter=50, return_singular_vectors=True)
return dot(diag(sqrt(s[:self.ell])), Vt[:self.ell,:])
