Guidelines ----------

To add extra doc to the operator `XXX`, write a class `XXXDoc`, deriving from the base class `SymbolDoc`, and put the extra doc as the docstring of `XXXDoc`.

The document added here should be Python-specific. Documents that are useful for all language bindings should be added to the C++ side where the operator is defined / registered.

The code snippet in the docstring will be run using `doctest`. During running, the environment will have access to

- all the global names in this file (e.g. `SymbolDoc`) - all the operators (e.g. `FullyConnected`) - the name `test_utils` for `mxnet.test_utils` (e.g. `test_utils.reldiff`) - the name `mxnet` (e.g. `mxnet.nd.zeros`) - the name `numpy`

The following documents are recommended:

- *Examples*: simple and short code snippet showing how to use this operator. It should show typical calling examples and behaviors (e.g. maps an input of what shape to an output of what shape). - *Regression Test*: longer test code for the operators. We normally do not expect the users to read those, but they will be executed by `doctest` to ensure the behavior of each operator does not change unintentionally.

get_output_shape

    Get user friendly information of the output shapes.

build_doc

    Build docstring for symbolic functions.

class ActivationDoc(SymbolDoc): """ Examples -------- A one-hidden-layer MLP with ReLU activation:

    >>> data = Variable('data')
    >>> mlp = FullyConnected(data=data, num_hidden=128, name='proj')
    >>> mlp = Activation(data=mlp, act_type='relu', name='activation')
    >>> mlp = FullyConnected(data=mlp, num_hidden=10, name='mlp')
    >>> mlp
    <Symbol mlp>

    Regression Test
    ---------------
    ReLU activation

    >>> test_suites = [
    ...     ('relu', lambda x: numpy.maximum(x, 0)),
    ...     ('sigmoid', lambda x: 1 / (1 + numpy.exp(-x))),
    ...     ('tanh', lambda x: numpy.tanh(x)),
    ...     ('softrelu', lambda x: numpy.log(1 + numpy.exp(x)))
    ... ]
    >>> x = test_utils.random_arrays((2, 3, 4))
    >>> for act_type, numpy_impl in test_suites:
    ...     op = Activation(act_type=act_type, name='act')
    ...     y = test_utils.simple_forward(op, act_data=x)
    ...     y_np = numpy_impl(x)
    ...     print('%s: %s' % (act_type, test_utils.almost_equal(y, y_np)))
    relu: True
    sigmoid: True
    tanh: True
    softrelu: True
    """

class DropoutDoc(SymbolDoc): """ Examples -------- Apply dropout to corrupt input as zero with probability 0.2:

    >>> data = Variable('data')
    >>> data_dp = Dropout(data=data, p=0.2)

    Regression Test
    ---------------
    >>> shape = (100, 100)  # take larger shapes to be more statistical stable
    >>> x = numpy.ones(shape)
    >>> op = Dropout(p=0.5, name='dp')
    >>> # dropout is identity during testing
    >>> y = test_utils.simple_forward(op, dp_data=x, is_train=False)
    >>> test_utils.almost_equal(x, y, threshold=0)
    True
    >>> y = test_utils.simple_forward(op, dp_data=x, is_train=True)
    >>> # expectation is (approximately) unchanged
    >>> numpy.abs(x.mean() - y.mean()) < 0.1
    True
    >>> set(numpy.unique(y)) == set([0, 2])
    True
    """

class EmbeddingDoc(SymbolDoc): """ Examples -------- Assume we want to map the 26 English alphabet letters to 16-dimensional vectorial representations.

    >>> vocabulary_size = 26
    >>> embed_dim = 16
    >>> seq_len, batch_size = (10, 64)
    >>> input = Variable('letters')
    >>> op = Embedding(data=input, input_dim=vocabulary_size, output_dim=embed_dim,
    ...                name='embed')
    >>> SymbolDoc.get_output_shape(op, letters=(seq_len, batch_size))
    {'embed_output': (10L, 64L, 16L)}

    Regression Test
    ---------------
    >>> vocab_size, embed_dim = (26, 16)
    >>> batch_size = 12
    >>> word_vecs = test_utils.random_arrays((vocab_size, embed_dim))
    >>> op = Embedding(name='embed', input_dim=vocab_size, output_dim=embed_dim)
    >>> x = numpy.random.choice(vocab_size, batch_size)
    >>> y = test_utils.simple_forward(op, embed_data=x, embed_weight=word_vecs)
    >>> y_np = word_vecs[x]
    >>> test_utils.almost_equal(y, y_np)
    True
    """

class FlattenDoc(SymbolDoc): """ Examples -------- Flatten is usually applied before `FullyConnected`, to reshape the 4D tensor produced by convolutional layers to 2D matrix:

    >>> data = Variable('data')  # say this is 4D from some conv/pool
    >>> flatten = Flatten(data=data, name='flat')  # now this is 2D
    >>> SymbolDoc.get_output_shape(flatten, data=(2, 3, 4, 5))
    {'flat_output': (2L, 60L)}

    Regression Test
    ---------------
    >>> test_dims = [(2, 3, 4, 5), (2, 3), (2,)]
    >>> op = Flatten(name='flat')
    >>> for dims in test_dims:
    ...     x = test_utils.random_arrays(dims)
    ...     y = test_utils.simple_forward(op, flat_data=x)
    ...     y_np = x.reshape((dims[0], numpy.prod(dims[1:])))
    ...     print('%s: %s' % (dims, test_utils.almost_equal(y, y_np)))
    (2, 3, 4, 5): True
    (2, 3): True
    (2,): True
    """

class FullyConnectedDoc(SymbolDoc): """ Examples -------- Construct a fully connected operator with target dimension 512.

    >>> data = Variable('data')  # or some constructed NN
    >>> op = FullyConnected(data=data,
    ...                     num_hidden=512,
    ...                     name='FC1')
    >>> op
    <Symbol FC1>
    >>> SymbolDoc.get_output_shape(op, data=(128, 100))
    {'FC1_output': (128L, 512L)}

    A simple 3-layer MLP with ReLU activation:

    >>> net = Variable('data')
    >>> for i, dim in enumerate([128, 64]):
    ...     net = FullyConnected(data=net, num_hidden=dim, name='FC%d' % i)
    ...     net = Activation(data=net, act_type='relu', name='ReLU%d' % i)
    >>> # 10-class predictor (e.g. MNIST)
    >>> net = FullyConnected(data=net, num_hidden=10, name='pred')
    >>> net
    <Symbol pred>

    Regression Test
    ---------------
    >>> dim_in, dim_out = (3, 4)
    >>> x, w, b = test_utils.random_arrays((10, dim_in), (dim_out, dim_in), (dim_out,))
    >>> op = FullyConnected(num_hidden=dim_out, name='FC')
    >>> out = test_utils.simple_forward(op, FC_data=x, FC_weight=w, FC_bias=b)
    >>> # numpy implementation of FullyConnected
    >>> out_np = numpy.dot(x, w.T) + b
    >>> test_utils.almost_equal(out, out_np)
    True
    """

class ConcatDoc(SymbolDoc): """ Examples -------- Concat two (or more) inputs along a specific dimension:

    >>> a = Variable('a')
    >>> b = Variable('b')
    >>> c = Concat(a, b, dim=1, name='my-concat')
    >>> c
    <Symbol my-concat>
    >>> SymbolDoc.get_output_shape(c, a=(128, 10, 3, 3), b=(128, 15, 3, 3))
    {'my-concat_output': (128L, 25L, 3L, 3L)}

    Note the shape should be the same except on the dimension that is being
    concatenated.
    """

class BroadcastPlusDoc(SymbolDoc): """ Examples --------

    >>> a = Variable('a')
    >>> b = Variable('b')
    >>> c = broadcast_plus(a, b)

    Normal summation with matching shapes:

    >>> dev = mxnet.context.cpu();
    >>> x = c.bind(dev, args={'a': mxnet.nd.ones((2, 2)), 'b' : mxnet.nd.ones((2, 2))})
    >>> x.forward()
    [<NDArray 2x2 @cpu(0)>]
    >>> print x.outputs[0].asnumpy()
    [[ 2.  2.]
     [ 2.  2.]]

    Broadcasting:

    >>> x = c.bind(dev, args={'a': mxnet.nd.ones((2, 2)), 'b' : mxnet.nd.ones((1, 1))})
    >>> x.forward()
    [<NDArray 2x2 @cpu(0)>]
    >>> print x.outputs[0].asnumpy()
    [[ 2.  2.]
     [ 2.  2.]]

    >>> x = c.bind(dev, args={'a': mxnet.nd.ones((2, 1)), 'b' : mxnet.nd.ones((1, 2))})
    >>> x.forward()
    [<NDArray 2x2 @cpu(0)>]
    >>> print x.outputs[0].asnumpy()
    [[ 2.  2.]
     [ 2.  2.]]

    >>> x = c.bind(dev, args={'a': mxnet.nd.ones((1, 2)), 'b' : mxnet.nd.ones((2, 1))})
    >>> x.forward()
    [<NDArray 2x2 @cpu(0)>]
    >>> print x.outputs[0].asnumpy()
    [[ 2.  2.]
     [ 2.  2.]]
    """