Python tensorflow.python.framework.ops.Operation() Examples

def testKFeatureTrainingConstruction(self):
    # pylint: disable=W0612
    data = constant_op.constant(
        [[random.uniform(-1, 1) for i in range(self.params.num_features)]
         for _ in range(100)])

    labels = [1 for _ in range(100)]

    with variable_scope.variable_scope(
        "KFeatureDecisionsToDataThenNNTest.testKFeatureTrainingContruction"):
      graph_builder = (
          k_feature_decisions_to_data_then_nn.KFeatureDecisionsToDataThenNN(
              self.params))
      graph = graph_builder.training_graph(data, labels, None)

      self.assertTrue(isinstance(graph, Operation)) 

def get_copied_op(org_instance, graph, scope=""):
  """Given an `Operation` instance from some `Graph`, returns
  its namesake from `graph`, under the specified scope
  (default `""`).

  If a copy of `org_instance` is present in `graph` under the given
  `scope`, it will be returned.

  Args:
  org_instance: An `Operation` from some `Graph`.
  graph: The `Graph` to be searched for a copr of `org_instance`.
  scope: The scope `org_instance` is present in.

  Returns:
      The `Operation` copy from `graph`.
  """

  #The name of the copied instance
  if scope != '':
    new_name = scope + '/' + org_instance.name
  else:
    new_name = org_instance.name

  return graph.as_graph_element(new_name, allow_tensor=True,
                                allow_operation=True) 

def reroute_ts(ts0, ts1, can_modify=None, cannot_modify=None):
  """For each tensor's pair, replace the end of t1 by the end of t0.

  B0 B1     B0 B1
  |  |    => |/
  A0 A1     A0 A1

  The end of the tensors in ts1 are left dangling.

  Args:
    ts0: an object convertible to a list of `tf.Tensor`.
    ts1: an object convertible to a list of `tf.Tensor`.
    can_modify: iterable of operations which can be modified. Any operation
      outside within_ops will be left untouched by this function.
    cannot_modify: iterable of operations which cannot be modified. Any
      operation within cannot_modify will be left untouched by this function.
  Returns:
    The number of individual modifications made by the function.
  Raises:
    TypeError: if ts0 or ts1 cannot be converted to a list of tf.Tensor.
    TypeError: if can_modify or cannot_modify is not None and cannot be
      converted to a list of tf.Operation.
  """
  return _reroute_ts(ts0, ts1, _RerouteMode.a2b, can_modify, cannot_modify) 

def _BuildCondTensor(self, v):
    if isinstance(v, ops.Operation):
      # Use pivot as the proxy for this op.
      return with_dependencies([v], self._pivot)
    elif isinstance(v, (ops.IndexedSlices, sparse_tensor.SparseTensor)):
      values = self._ProcessOutputTensor(v.values)
      indices = self._ProcessOutputTensor(v.indices)
      if isinstance(v, ops.IndexedSlices):
        dense_shape = v.dense_shape
        if dense_shape is not None:
          dense_shape = self._ProcessOutputTensor(dense_shape)
        return ops.IndexedSlices(values, indices, dense_shape)
      else:
        dense_shape = self._ProcessOutputTensor(v.dense_shape)
        return sparse_tensor.SparseTensor(indices, values, dense_shape)
    else:
      v = nest.map_structure(_convert_tensorarray_to_flow, v)
      return self._ProcessOutputTensor(ops.convert_to_tensor(v)) 

def _get_fetch_names(fetches):
  """Get a flattened list of the names in run() call fetches.

  Args:
    fetches: Fetches of the `Session.run()` call. It maybe a Tensor, an
      Operation or a Variable. It may also be nested lists, tuples or
      dicts. See doc of `Session.run()` for more details.

  Returns:
    (list of str) A flattened list of fetch names from `fetches`.
  """

  lines = []
  if isinstance(fetches, (list, tuple)):
    for fetch in fetches:
      lines.extend(_get_fetch_names(fetch))
  elif isinstance(fetches, dict):
    for key in fetches:
      lines.extend(_get_fetch_names(fetches[key]))
  else:
    # This ought to be a Tensor, an Operation or a Variable, for which the name
    # attribute should be available. (Bottom-out condition of the recursion.)
    lines.append(_get_fetch_name(fetches))

  return lines 

def is_placeholder(self, graph_element_name):
    """Check whether a graph element is a Placeholder, by name.

    Args:
      graph_element_name: (str) Name of the tensor or op to be tested.

    Returns:
      (bool) Whether the graph element of the specified name is a Placeholder
        op or the output Tensor of a Placeholder op.

    Raises:
      ValueError: If graph_element_name is not in the transitive closure of the
        stepper instance.
    """

    node_name = self._get_node_name(graph_element_name)
    if node_name not in self.sorted_nodes():
      raise ValueError(
          "%s is not in the transitive closure of this NodeStepper "
          "instance" % graph_element_name)

    graph_element = self._sess.graph.as_graph_element(graph_element_name)
    if not isinstance(graph_element, ops.Operation):
      graph_element = graph_element.op
    return graph_element.type == "Placeholder" 

def swap_ts(ts0, ts1, can_modify=None, cannot_modify=None):
  """For each tensor's pair, swap the end of (t0,t1).

  B0 B1     B0 B1
  |  |    =>  X
  A0 A1     A0 A1

  Args:
    ts0: an object convertible to a list of `tf.Tensor`.
    ts1: an object convertible to a list of `tf.Tensor`.
    can_modify: iterable of operations which can be modified. Any operation
      outside within_ops will be left untouched by this function.
    cannot_modify: iterable of operations which cannot be modified.
      Any operation within cannot_modify will be left untouched by this
      function.
  Returns:
    The number of individual modifications made by the function.
  Raises:
    TypeError: if ts0 or ts1 cannot be converted to a list of tf.Tensor.
    TypeError: if can_modify or cannot_modify is not None and cannot be
      converted to a list of tf.Operation.
  """
  return _reroute_ts(ts0, ts1, _RerouteMode.swap, can_modify, cannot_modify) 

def remove_control_inputs(op, cops):
  """Remove the control inputs cops from co.

  Warning: this function is directly manipulating the internals of the
  `tf.Graph`.

  Args:
    op: a `tf.Operation` from which to remove the control inputs.
    cops: an object convertible to a list of `tf.Operation`.
  Raises:
    TypeError: if op is not a `tf.Operation`.
    ValueError: if any cop in cops is not a control input of op.
  """
  if not isinstance(op, _tf_ops.Operation):
    raise TypeError("Expected a tf.Operation, got: {}", type(op))
  cops = _util.make_list_of_op(cops, allow_graph=False)
  for cop in cops:
    if cop not in op.control_inputs:
      raise ValueError("{} is not a control_input of {}".format(op.name,
                                                                cop.name))
  # pylint: disable=protected-access
  op._control_inputs = [cop for cop in op._control_inputs if cop not in cops]
  op._recompute_node_def()
  # pylint: enable=protected-access 

def add_control_inputs(op, cops):
  """Add the control inputs cops to op.

  Warning: this function is directly manipulating the internals of the tf.Graph.

  Args:
    op: a tf.Operation to which the control inputs are added.
    cops: an object convertible to a list of `tf.Operation`.
  Raises:
    TypeError: if op is not a tf.Operation
    ValueError: if any cop in cops is already a control input of op.
  """
  if not isinstance(op, _tf_ops.Operation):
    raise TypeError("Expected a tf.Operation, got: {}", type(op))
  cops = _util.make_list_of_op(cops, allow_graph=False)
  for cop in cops:
    if cop in op.control_inputs:
      raise ValueError("{} is already a control_input of {}".format(cop.name,
                                                                    op.name))
  # pylint: disable=protected-access
  op._control_inputs += cops
  op._recompute_node_def()
  # pylint: enable=protected-access 

def find_corresponding(targets, dst_graph, dst_scope="", src_scope=""):
  """Find corresponding ops/tensors in a different graph.

  `targets` is a Python tree, that is, a nested structure of iterable
  (list, tupple, dictionary) whose leaves are instances of
  `tf.Tensor` or `tf.Operation`

  Args:
    targets: A Python tree containing `tf.Tensor` or `tf.Operation`
      belonging to the original graph.
    dst_graph: The graph in which the corresponding graph element must be found.
    dst_scope: A scope which is prepended to the name to look for.
    src_scope: A scope which is removed from the original of `top` name.

  Returns:
    A Python tree containin the corresponding tf.Tensor` or a `tf.Operation`.

  Raises:
    ValueError: if `src_name` does not start with `src_scope`.
    TypeError: if `top` is not a `tf.Tensor` or a `tf.Operation`
    KeyError: If the corresponding graph element cannot be found.
  """
  def func(top):
    return find_corresponding_elem(top, dst_graph, dst_scope, src_scope)
  return transform_tree(targets, func) 

def get_consuming_ops(ts):
  """Return all the consuming ops of the tensors in ts.

  Args:
    ts: a list of `tf.Tensor`
  Returns:
    A list of all the consuming `tf.Operation` of the tensors in `ts`.
  Raises:
    TypeError: if ts cannot be converted to a list of `tf.Tensor`.
  """
  ts = make_list_of_t(ts, allow_graph=False)
  ops = []
  for t in ts:
    for op in t.consumers():
      if op not in ops:
        ops.append(op)
  return ops 

def get_consuming_ops(ts):
  """Return all the consuming ops of the tensors in ts.

  Args:
    ts: a list of `tf.Tensor`
  Returns:
    A list of all the consuming `tf.Operation` of the tensors in `ts`.
  Raises:
    TypeError: if ts cannot be converted to a list of `tf.Tensor`.
  """
  ts = make_list_of_t(ts, allow_graph=False)
  ops = []
  for t in ts:
    for op in t.consumers():
      if op not in ops:
        ops.append(op)
  return ops 

def find_corresponding(targets, dst_graph, dst_scope="", src_scope=""):
  """Find corresponding ops/tensors in a different graph.

  `targets` is a Python tree, that is, a nested structure of iterable
  (list, tupple, dictionary) whose leaves are instances of
  `tf.Tensor` or `tf.Operation`

  Args:
    targets: A Python tree containing `tf.Tensor` or `tf.Operation`
      belonging to the original graph.
    dst_graph: The graph in which the corresponding graph element must be found.
    dst_scope: A scope which is prepended to the name to look for.
    src_scope: A scope which is removed from the original of `top` name.

  Returns:
    A Python tree containin the corresponding tf.Tensor` or a `tf.Operation`.

  Raises:
    ValueError: if `src_name` does not start with `src_scope`.
    TypeError: if `top` is not a `tf.Tensor` or a `tf.Operation`
    KeyError: If the corresponding graph element cannot be found.
  """
  def func(top):
    return find_corresponding_elem(top, dst_graph, dst_scope, src_scope)
  return transform_tree(targets, func) 

def __init__(self, graph):
    """Create a dictionary of control-output dependencies.

    Args:
      graph: a `tf.Graph`.
    Returns:
      A dictionary where a key is a `tf.Operation` instance and the
         corresponding value is a list of all the ops which have the key
         as one of their control-input dependencies.
    Raises:
      TypeError: graph is not a `tf.Graph`.
    """
    if not isinstance(graph, tf_ops.Graph):
      raise TypeError("Expected a tf.Graph, got: {}".format(type(graph)))
    self._control_outputs = {}
    self._graph = graph
    self._version = None
    self._build() 

def testTrainingConstructionRegression(self):
    input_data = [[-1., 0.], [-1., 2.],  # node 1
                  [1., 0.], [1., -2.]]  # node 2
    input_labels = [0, 1, 2, 3]

    params = tensor_forest.ForestHParams(
        num_classes=4,
        num_features=2,
        num_trees=10,
        max_nodes=1000,
        split_after_samples=25,
        regression=True).fill()

    graph_builder = tensor_forest.RandomForestGraphs(params)
    graph = graph_builder.training_graph(input_data, input_labels)
    self.assertTrue(isinstance(graph, ops.Operation)) 

def testTrainingConstructionClassificationSparse(self):
    input_data = sparse_tensor.SparseTensor(
        indices=[[0, 0], [0, 3], [1, 0], [1, 7], [2, 1], [3, 9]],
        values=[-1.0, 0.0, -1., 2., 1., -2.0],
        dense_shape=[4, 10])
    input_labels = [0, 1, 2, 3]

    params = tensor_forest.ForestHParams(
        num_classes=4,
        num_features=10,
        num_trees=10,
        max_nodes=1000,
        split_after_samples=25).fill()

    graph_builder = tensor_forest.RandomForestGraphs(params)
    graph = graph_builder.training_graph(input_data, input_labels)
    self.assertTrue(isinstance(graph, ops.Operation)) 

def swap_ts(ts0, ts1, can_modify=None, cannot_modify=None):
  """For each tensor's pair, swap the end of (t0,t1).

  B0 B1     B0 B1
  |  |    =>  X
  A0 A1     A0 A1

  Args:
    ts0: an object convertible to a list of `tf.Tensor`.
    ts1: an object convertible to a list of `tf.Tensor`.
    can_modify: iterable of operations which can be modified. Any operation
      outside within_ops will be left untouched by this function.
    cannot_modify: iterable of operations which cannot be modified.
      Any operation within cannot_modify will be left untouched by this
      function.
  Returns:
    The number of individual modifications made by the function.
  Raises:
    TypeError: if ts0 or ts1 cannot be converted to a list of tf.Tensor.
    TypeError: if can_modify or cannot_modify is not None and cannot be
      converted to a list of tf.Operation.
  """
  return _reroute_ts(ts0, ts1, _RerouteMode.swap, can_modify, cannot_modify) 

def reroute_ts(ts0, ts1, can_modify=None, cannot_modify=None):
  """For each tensor's pair, replace the end of t1 by the end of t0.

  B0 B1     B0 B1
  |  |    => |/
  A0 A1     A0 A1

  The end of the tensors in ts1 are left dangling.

  Args:
    ts0: an object convertible to a list of `tf.Tensor`.
    ts1: an object convertible to a list of `tf.Tensor`.
    can_modify: iterable of operations which can be modified. Any operation
      outside within_ops will be left untouched by this function.
    cannot_modify: iterable of operations which cannot be modified. Any
      operation within cannot_modify will be left untouched by this function.
  Returns:
    The number of individual modifications made by the function.
  Raises:
    TypeError: if ts0 or ts1 cannot be converted to a list of tf.Tensor.
    TypeError: if can_modify or cannot_modify is not None and cannot be
      converted to a list of tf.Operation.
  """
  return _reroute_ts(ts0, ts1, _RerouteMode.a2b, can_modify, cannot_modify) 

def remove_control_inputs(op, cops):
  """Remove the control inputs cops from co.

  Warning: this function is directly manipulating the internals of the
  `tf.Graph`.

  Args:
    op: a `tf.Operation` from which to remove the control inputs.
    cops: an object convertible to a list of `tf.Operation`.
  Raises:
    TypeError: if op is not a `tf.Operation`.
    ValueError: if any cop in cops is not a control input of op.
  """
  if not isinstance(op, tf_ops.Operation):
    raise TypeError("Expected a tf.Operation, got: {}", type(op))
  cops = util.make_list_of_op(cops, allow_graph=False)
  for cop in cops:
    if cop not in op.control_inputs:
      raise ValueError("{} is not a control_input of {}".format(op.name,
                                                                cop.name))
  # pylint: disable=protected-access
  op._control_inputs = [cop for cop in op._control_inputs if cop not in cops]
  op._recompute_node_def()
  # pylint: enable=protected-access 

def add_control_inputs(op, cops):
  """Add the control inputs cops to co.

  Warning: this function is directly manipulating the internals of the tf.Graph.

  Args:
    op: a tf.Operation to which the control inputs are added.
    cops: an object convertible to a list of `tf.Operation`.
  Raises:
    TypeError: if op is not a tf.Operation
    ValueError: if any cop in cops is already a control input of op.
  """
  if not isinstance(op, tf_ops.Operation):
    raise TypeError("Expected a tf.Operation, got: {}", type(op))
  cops = util.make_list_of_op(cops, allow_graph=False)
  for cop in cops:
    if cop in op.control_inputs:
      raise ValueError("{} is already a control_input of {}".format(op.name,
                                                                    cop.name))
  # pylint: disable=protected-access
  op._control_inputs += cops
  op._recompute_node_def()
  # pylint: enable=protected-access 

def assign_renamed_collections_handler(info, elem, elem_):
  """Add the transformed elem to the (renamed) collections of elem.

  A collection is renamed only if is not a known key, as described in
  `tf.GraphKeys`.

  Args:
    info: Transform._TmpInfo instance.
    elem: the original element (`tf.Tensor` or `tf.Operation`)
    elem_: the transformed element
  """
  known_collection_names = util.get_predefined_collection_names()
  for name, collection in iteritems(info.collections):
    if elem not in collection:
      continue

    if name in known_collection_names:
      transformed_name = name
    else:
      transformed_name = info.new_name(name)
    info.graph_.add_to_collection(transformed_name, elem_) 

def _get_fetch_names(fetches):
  """Get a flattened list of the names in run() call fetches.

  Args:
    fetches: Fetches of the `Session.run()` call. It maybe a Tensor, an
      Operation or a Variable. It may also be nested lists, tuples or
      dicts. See doc of `Session.run()` for more details.

  Returns:
    (list of str) A flattened list of fetch names from `fetches`.
  """

  lines = []
  if isinstance(fetches, (list, tuple)):
    for fetch in fetches:
      lines.extend(_get_fetch_names(fetch))
  elif isinstance(fetches, dict):
    for key in fetches:
      lines.extend(_get_fetch_names(fetches[key]))
  else:
    # This ought to be a Tensor, an Operation or a Variable, for which the name
    # attribute should be available. (Bottom-out condition of the recursion.)
    lines.append(_get_fetch_name(fetches))

  return lines 

def is_placeholder(self, graph_element_name):
    """Check whether a graph element is a Placeholder, by name.

    Args:
      graph_element_name: (str) Name of the tensor or op to be tested.

    Returns:
      (bool) Whether the graph element of the specified name is a Placeholder
        op or the output Tensor of a Placeholder op.

    Raises:
      ValueError: If graph_element_name is not in the transitive closure of the
        stepper instance.
    """

    node_name = self._get_node_name(graph_element_name)
    if node_name not in self.sorted_nodes():
      raise ValueError(
          "%s is not in the transitive closure of this NodeStepper "
          "instance" % graph_element_name)

    graph_element = self._sess.graph.as_graph_element(graph_element_name)
    if not isinstance(graph_element, ops.Operation):
      graph_element = graph_element.op
    return graph_element.type == "Placeholder" 

def __init__(self, graph):
    """Create a dictionary of control-output dependencies.

    Args:
      graph: a `tf.Graph`.
    Returns:
      A dictionary where a key is a `tf.Operation` instance and the
         corresponding value is a list of all the ops which have the key
         as one of their control-input dependencies.
    Raises:
      TypeError: graph is not a `tf.Graph`.
    """
    if not isinstance(graph, tf_ops.Graph):
      raise TypeError("Expected a tf.Graph, got: {}".format(type(graph)))
    self._control_outputs = {}
    self._graph = graph
    self._version = None
    self._build() 

def build_results(self, values):
    if not values:
      # 'Operation' case
      return None
    else:
      return self._contraction_fn(values) 

def _get_transformed_map(self, top):
    """Return the correct container depending on the type of `top`."""
    if isinstance(top, tf_ops.Operation):
      return self._transformed_ops
    elif isinstance(top, tf_ops.Tensor):
      return self._transformed_ts
    else:
      raise TypeError(
          "Expected a tf.Tensor or a tf.Operation, got a {}".format(
              type(top))) 

def find_corresponding_elem(target, dst_graph, dst_scope="", src_scope=""):
  """Find corresponding op/tensor in a different graph.

  Args:
    target: A `tf.Tensor` or a `tf.Operation` belonging to the original graph.
    dst_graph: The graph in which the corresponding graph element must be found.
    dst_scope: A scope which is prepended to the name to look for.
    src_scope: A scope which is removed from the original of `target` name.

  Returns:
    The corresponding tf.Tensor` or a `tf.Operation`.

  Raises:
    ValueError: if `src_name` does not start with `src_scope`.
    TypeError: if `target` is not a `tf.Tensor` or a `tf.Operation`
    KeyError: If the corresponding graph element cannot be found.
  """
  src_name = target.name
  if src_scope:
    src_scope = scope_finalize(src_scope)
    if not src_name.startswidth(src_scope):
      raise ValueError("{} does not start with {}".format(src_name, src_scope))
    src_name = src_name[len(src_scope):]

  dst_name = src_name
  if dst_scope:
    dst_scope = scope_finalize(dst_scope)
    dst_name = dst_scope + dst_name

  if isinstance(target, tf_ops.Tensor):
    return dst_graph.get_tensor_by_name(dst_name)
  if isinstance(target, tf_ops.Operation):
    return dst_graph.get_operation_by_name(dst_name)
  raise TypeError("Expected tf.Tensor or tf.Operation, got: {}", type(target)) 

def __init__(self, graph, fetches, feeds):
    """Creates a fetch handler.

    Args:
      graph: Graph of the fetches.   Used to check for fetchability
        and to convert all fetches to tensors or ops as needed.
      fetches: An arbitrary fetch structure: singleton, list, tuple,
        namedtuple, or dict.
      feeds: A feed dict where keys are fully resolved tensor names.
    """
    with graph.as_default():
      self._fetch_mapper = _FetchMapper.for_fetch(fetches)
    self._fetches = []
    self._targets = []
    self._feeds = feeds
    self._ops = []
    self._fetch_handles = {}
    for fetch in self._fetch_mapper.unique_fetches():
      fetch_name = compat.as_bytes(fetch.name)
      if isinstance(fetch, ops.Operation):
        self._assert_fetchable(graph, fetch)
        self._targets.append(fetch_name)
        self._ops.append(True)
      else:
        self._assert_fetchable(graph, fetch.op)
        self._fetches.append(fetch_name)
        self._ops.append(False)
      # Remember the fetch if it is for a tensor handle.
      if isinstance(fetch, ops.Tensor) and fetch.op.type == 'GetSessionHandle':
        self._fetch_handles[fetch_name] = fetch.op.inputs[0].dtype
    self._final_fetches = [x for x in self._fetches if x not in feeds] 

def _WeMustGoDeeper(self, msg):
    with self.assertRaisesOpError(msg):
      node_def = ops._NodeDef("op_type", "name")
      node_def_orig = ops._NodeDef("op_type_orig", "orig")
      op_orig = ops.Operation(node_def_orig, ops.get_default_graph())
      op = ops.Operation(node_def, ops.get_default_graph(), original_op=op_orig)
      raise errors.UnauthenticatedError(node_def, op, "true_err") 

def is_feedable(self, name):
    """Determine if a graph element if feedable.

    Args:
      name: (str) name of the graph element (Tensor or Operation)

    Returns:
      (bool) whether the graph element is feedable.
    """

    if not isinstance(name, six.string_types):
      raise TypeError("Expected type str; got type %s" % type(name))

    elem = self._sess.graph.as_graph_element(name)
    return self._sess.graph.is_feedable(elem)