def transform(self, X, copy=True):
"""Transform a count matrix to a tf or tf-idf representation
Parameters
----------
X : array-like of (n_samples, n_features)
A matrix of term/token counts
copy : bool, default=True
Whether to copy X and operate on the copy or perform in-place
operations.
Returns
-------
vectors : array-like of shape (n_samples, n_features)
"""
if copy:
X = X.copy()
dtype = _get_dtype(X)
X = self._convert_to_csr(X, dtype)
if X.dtype != dtype:
X = X.astype(dtype)
n_samples, n_features = X.shape
if self.sublinear_tf:
cp.log(X.data, X.data)
X.data += 1
if self.use_idf:
self._check_is_idf_fitted()
expected_n_features = self._idf_diag.shape[0]
if n_features != expected_n_features:
raise ValueError("Input has n_features=%d while the model"
" has been trained with n_features=%d" %
(n_features, expected_n_features))
csr_diag_mul(X, self._idf_diag, inplace=True)
if self.norm:
if self.norm == 'l1':
csr_row_normalize_l1(X, inplace=True)
elif self.norm == 'l2':
csr_row_normalize_l2(X, inplace=True)
return X
def transform(self, raw_documents):
"""
Transform documents to document-term matrix.
Extract token counts out of raw text documents using the vocabulary
fitted with fit or the one provided to the constructor.
Parameters
----------
raw_documents : cudf.Series
A Series of string documents
Returns
-------
X : sparse CuPy CSR matrix of shape (n_samples, n_features)
Document-term matrix.
"""
docs = self._preprocess(raw_documents)
del raw_documents
n_doc = len(docs)
tokenized_df = self._create_tokenized_df(docs)
del docs
count_df = self._count_hash(tokenized_df)
del tokenized_df
empty_doc_ids = self._compute_empty_doc_ids(count_df, n_doc)
X = create_csr_matrix_from_count_df(count_df,
empty_doc_ids,
n_doc,
self.n_features,
dtype=self.dtype)
if self.binary:
X.data.fill(1)
if self.norm:
if self.norm == "l1":
csr_row_normalize_l1(X, inplace=True)
elif self.norm == "l2":
csr_row_normalize_l2(X, inplace=True)
return X