dict to scipy.sparse

dict2sparse.py

import numpy as np
from scipy.sparse import csr_matrix

def dict2sparse(d):
    data = list(d.values())
    indices = list(d.keys())
    indptr = [0, len(d)]
    return csr_matrix((data, indices, indptr), shape=(1, max(d)+1), dtype=np.uint32)


"""
> d = {1: 10, 2: 20, 3: 30}
> X = dict2sparse(d)
> X.toarray()
# => array([[ 0, 10, 20, 30]], dtype=uint32)
"""

dict2sparse_compare.py

import numpy as np
from scipy.sparse import csr_matrix


def counter_to_sparse1(counter):
    vec = csr_matrix((1, max(counter)+1), dtype=np.uint32)
    for (k, v) in counter.items():
        vec[0, k] = v
    return vec

def counter_to_sparse2(counter):
    data = list(counter.values())
    indices = list(counter.keys())
    indptr = [0, len(counter)]

    X = csr_matrix((data, indices, indptr), shape=(1, max(d)+1), dtype=np.uint32)
    return X