B Zvd$(@sdZddZe\ZZddlmZddlmZddlm Z ddlm Z dd l m Z dd lmZdd lmZdd lmZGd ddfeZeedddddddddZ dS)z?Implement generators for ``keras`` which will balance the data.cCsNdd}dd}|\}}|\}}|p*|}|r@|r:|}qF|}ntf}||fS)zzTry to import keras from keras and tensorflow. This is possible to import the sequence from keras or tensorflow. cSsVy6ddl}t|jdr$|jjfdfS|jjjfdfSWntk rPtdfSXdS)NSequenceTF)kerashasattrutilsr data_utils ImportErrortuple)rr K/var/www/html/venv/lib/python3.7/site-packages/imblearn/keras/_generator.pyimport_from_kerass z'import_keras..import_from_kerascSsZy:ddlm}t|jdr(|jjfdfS|jjjfdfSWntk rTtdfSXdS)Nr)rrTF) tensorflowrrrrrrr)rr r r import_from_tensforflows  z-import_keras..import_from_tensforflow)object)r r ZParentClassKerasZ has_keras_kZParentClassTensorflowZ has_keras_tfZ has_keras ParentClassr r r import_keras s   rr)issparse)clone)_safe_indexing)check_random_state)balanced_batch_generator)RandomUnderSampler) Substitution)_random_state_docstringc@sBeZdZdZdZddddddddZd d Zd d Zd dZdS)BalancedBatchGeneratora Create balanced batches when training a keras model. Create a keras ``Sequence`` which is given to ``fit``. The sampler defines the sampling strategy used to balance the dataset ahead of creating the batch. The sampler should have an attribute ``sample_indices_``. .. versionadded:: 0.4 Parameters ---------- X : ndarray of shape (n_samples, n_features) Original imbalanced dataset. y : ndarray of shape (n_samples,) or (n_samples, n_classes) Associated targets. sample_weight : ndarray of shape (n_samples,) Sample weight. sampler : sampler object, default=None A sampler instance which has an attribute ``sample_indices_``. By default, the sampler used is a :class:`~imblearn.under_sampling.RandomUnderSampler`. batch_size : int, default=32 Number of samples per gradient update. keep_sparse : bool, default=False Either or not to conserve or not the sparsity of the input (i.e. ``X``, ``y``, ``sample_weight``). By default, the returned batches will be dense. random_state : int, RandomState instance or None, default=None Control the randomization of the algorithm: - If int, ``random_state`` is the seed used by the random number generator; - If ``RandomState`` instance, random_state is the random number generator; - If ``None``, the random number generator is the ``RandomState`` instance used by ``np.random``. Attributes ---------- sampler_ : sampler object The sampler used to balance the dataset. indices_ : ndarray of shape (n_samples, n_features) The indices of the samples selected during sampling. Examples -------- >>> from sklearn.datasets import load_iris >>> iris = load_iris() >>> from imblearn.datasets import make_imbalance >>> class_dict = dict() >>> class_dict[0] = 30; class_dict[1] = 50; class_dict[2] = 40 >>> X, y = make_imbalance(iris.data, iris.target, sampling_strategy=class_dict) >>> import tensorflow >>> y = tensorflow.keras.utils.to_categorical(y, 3) >>> model = tensorflow.keras.models.Sequential() >>> model.add( ... tensorflow.keras.layers.Dense( ... y.shape[1], input_dim=X.shape[1], activation='softmax' ... ) ... ) >>> model.compile(optimizer='sgd', loss='categorical_crossentropy', ... metrics=['accuracy']) >>> from imblearn.keras import BalancedBatchGenerator >>> from imblearn.under_sampling import NearMiss >>> training_generator = BalancedBatchGenerator( ... X, y, sampler=NearMiss(), batch_size=10, random_state=42) >>> callback_history = model.fit(training_generator, epochs=10, verbose=0) TN F) sample_weightsampler batch_size keep_sparse random_statecCsBts td||_||_||_||_||_||_||_| dS)Nz'No module named 'keras') HAS_KERASrXyrrrrr _sample)selfr"r#rrrrr r r r __init__s zBalancedBatchGenerator.__init__cCsnt|j}|jdkr"t|d|_n t|j|_|j|j|jt |jdsTt d|jj |_ | |j dS)N)r sample_indices_z7'sampler' needs to have an attribute 'sample_indices_'.)rr rrZsampler_rZ fit_resampler"r#r ValueErrorr'indices_shuffle)r%r r r r r$s     zBalancedBatchGenerator._samplecCst|jj|jS)N)intr)sizer)r%r r r __len__szBalancedBatchGenerator.__len__cCst|j|j||j|d|j}t|j|j||j|d|j}t|rb|jsb|}|jdk rt|j|j||j|d|j}|jdkr||fS|||fSdS)N) rr"r)rr#rrZtoarrayr)r%indexZ X_resampledZ y_resampledZsample_weight_resampledr r r __getitem__s     z"BalancedBatchGenerator.__getitem__) __name__ __module__ __qualname____doc__Zuse_sequence_apir&r$r-r0r r r r r@sK  r)r NrF)rrrrr c Cst|||||||dS)a Create a balanced batch generator to train keras model. Returns a generator --- as well as the number of step per epoch --- which is given to ``fit``. The sampler defines the sampling strategy used to balance the dataset ahead of creating the batch. The sampler should have an attribute ``sample_indices_``. Parameters ---------- X : ndarray of shape (n_samples, n_features) Original imbalanced dataset. y : ndarray of shape (n_samples,) or (n_samples, n_classes) Associated targets. sample_weight : ndarray of shape (n_samples,), default=None Sample weight. sampler : sampler object, default=None A sampler instance which has an attribute ``sample_indices_``. By default, the sampler used is a :class:`~imblearn.under_sampling.RandomUnderSampler`. batch_size : int, default=32 Number of samples per gradient update. keep_sparse : bool, default=False Either or not to conserve or not the sparsity of the input (i.e. ``X``, ``y``, ``sample_weight``). By default, the returned batches will be dense. {random_state} Returns ------- generator : generator of tuple Generate batch of data. The tuple generated are either (X_batch, y_batch) or (X_batch, y_batch, sampler_weight_batch). steps_per_epoch : int The number of samples per epoch. Required by ``fit_generator`` in keras. Examples -------- >>> from sklearn.datasets import load_iris >>> X, y = load_iris(return_X_y=True) >>> from imblearn.datasets import make_imbalance >>> class_dict = dict() >>> class_dict[0] = 30; class_dict[1] = 50; class_dict[2] = 40 >>> from imblearn.datasets import make_imbalance >>> X, y = make_imbalance(X, y, sampling_strategy=class_dict) >>> import tensorflow >>> y = tensorflow.keras.utils.to_categorical(y, 3) >>> model = tensorflow.keras.models.Sequential() >>> model.add( ... tensorflow.keras.layers.Dense( ... y.shape[1], input_dim=X.shape[1], activation='softmax' ... ) ... ) >>> model.compile(optimizer='sgd', loss='categorical_crossentropy', ... metrics=['accuracy']) >>> from imblearn.keras import balanced_batch_generator >>> from imblearn.under_sampling import NearMiss >>> training_generator, steps_per_epoch = balanced_batch_generator( ... X, y, sampler=NearMiss(), batch_size=10, random_state=42) >>> callback_history = model.fit(training_generator, ... steps_per_epoch=steps_per_epoch, ... epochs=10, verbose=0) )r"r#rrrrr )tf_bbg)r"r#rrrrr r r r rsRr)r4rrr!Z scipy.sparserZ sklearn.baserZ sklearn.utilsrrr rr5Zunder_samplingrrrZutils._docstringrrr r r r s$ )