import numpy as np
from sklearn.decomposition import IncrementalPCA
from scipy import sparse

pca_op = IncrementalPCA(batch_size=10)
X = np.random.poisson(0.2, [100, 100])
for m in [sparse.csc_matrix, sparse.csr_matrix, sparse.dok_matrix, sparse.lil_matrix]:
    pca_op.fit_transform(m(X))

Traceback (most recent call last):
  File "<stdin>", line 2, in <module>
  File "/home/scottgigante/.local/lib/python3.5/site-packages/sklearn/base.py", line 464, in fit_transform
    return self.fit(X, **fit_params).transform(X)
  File "/home/scottgigante/.local/lib/python3.5/site-packages/sklearn/decomposition/incremental_pca.py", line 191, in fit
    X = check_array(X, copy=self.copy, dtype=[np.float64, np.float32])
  File "/home/scottgigante/.local/lib/python3.5/site-packages/sklearn/utils/validation.py", line 517, in check_array
    accept_large_sparse=accept_large_sparse)
  File "/home/scottgigante/.local/lib/python3.5/site-packages/sklearn/utils/validation.py", line 318, in _ensure_sparse_format
    raise TypeError('A sparse matrix was passed, but dense '
TypeError: A sparse matrix was passed, but dense data is required. Use X.toarray() to convert to a dense numpy array.

import numpy as np
from sklearn.decomposition import IncrementalPCA
from sklearn.utils import check_array, gen_batches
from scipy import sparse


class IncrementalPCA(IncrementalPCA):

    def fit(self, X, y=None):
        self.components_ = None
        self.n_samples_seen_ = 0
        self.mean_ = .0
        self.var_ = .0
        self.singular_values_ = None
        self.explained_variance_ = None
        self.explained_variance_ratio_ = None
        self.singular_values_ = None
        self.noise_variance_ = None
        X = check_array(X, accept_sparse=['csr', 'csc', 'dok', 'lil'], copy=self.copy, dtype=[np.float64, np.float32])
        n_samples, n_features = X.shape
        
        if self.batch_size is None:
            self.batch_size_ = 5 * n_features
        else:
            self.batch_size_ = self.batch_size
        for batch in gen_batches(n_samples, self.batch_size_,
                                 min_batch_size=self.n_components or 0):
            self.partial_fit(X[batch], check_input=True)
        return self

    def partial_fit(self, X, y=None, check_input=True):
        if check_input and sparse.issparse(X):
                X = X.toarray()
        super().partial_fit(X, y=y, check_input=check_input)

    def transform(self, X):
        n_samples = X.shape[0]
        output = []
        for batch in gen_batches(n_samples, self.batch_size_,
                                 min_batch_size=self.n_components or 0):
            X_batch = X[batch]
            if sparse.issparse(X_batch):
                X_batch = X_batch.toarray()
            output.append(super().transform(X_batch))
        return np.vstack(output)


pca_op = IncrementalPCA(batch_size=10)
X = np.random.poisson(0.2, [100, 100])
for m in [sparse.csc_matrix, sparse.csr_matrix, sparse.dok_matrix, sparse.lil_matrix]:
    pca_op.fit_transform(m(X))