common.py

# coding=utf-8
# Copyright 2020 The TF-Agents Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     https://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

"""Common utilities for TF-Agents."""

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import collections as cs
import contextlib
import distutils.version
import functools
import importlib
import os
from typing import Dict, Optional, Text

from absl import logging

import numpy as np
import tensorflow as tf

from tf_agents.specs import tensor_spec
from tf_agents.trajectories import time_step as ts
from tf_agents.typing import types
from tf_agents.utils import nest_utils
from tf_agents.utils import object_identity

# pylint:disable=g-direct-tensorflow-import
from tensorflow.core.protobuf import struct_pb2  # TF internal
from tensorflow.python import tf2 as tf2_checker  # TF internal
from tensorflow.python.eager import monitoring  # TF internal
from tensorflow.python.saved_model import nested_structure_coder  # TF internal
# pylint:enable=g-direct-tensorflow-import

try:
  importlib.import_module('tf_agents.utils.allow_tf1')
except ImportError:
  _TF1_MODE_ALLOWED = False
else:
  _TF1_MODE_ALLOWED = True


tf_agents_gauge = monitoring.BoolGauge('/tensorflow/agents/agents',
                                       'TF-Agents usage', 'method')


MISSING_RESOURCE_VARIABLES_ERROR = """
Resource variables are not enabled.  Please enable them by adding the following
code to your main() method:
  tf.compat.v1.enable_resource_variables()
For unit tests, subclass `tf_agents.utils.test_utils.TestCase`.
"""


def check_tf1_allowed():
  """Raises an error if running in TF1 (non-eager) mode and this is disabled."""
  if _TF1_MODE_ALLOWED:
    return
  if not tf2_checker.enabled():
    raise RuntimeError(
        'You are using TF1 or running TF with eager mode disabled.  '
        'TF-Agents no longer supports TF1 mode (except for a shrinking list of '
        'internal allowed users).  If this negatively affects you, please '
        'reach out to the TF-Agents team.  Otherwise please use TF2.')


def resource_variables_enabled():
  return tf.compat.v1.resource_variables_enabled()


_IN_LEGACY_TF1 = (
    tf.__git_version__ != 'unknown'
    and tf.__version__ != '1.15.0'
    and (distutils.version.LooseVersion(tf.__version__) <=
         distutils.version.LooseVersion('1.15.0.dev20190821')))


def in_legacy_tf1():
  return _IN_LEGACY_TF1


def set_default_tf_function_parameters(*args, **kwargs):
  """Generates a decorator that sets default parameters for `tf.function`.

  Args:
    *args: default arguments for the `tf.function`.
    **kwargs: default keyword arguments for the `tf.function`.

  Returns:
    Function decorator with preconfigured defaults for `tf.function`.
  """
  def maybe_wrap(fn):
    """Helper function."""
    wrapped = [None]

    @functools.wraps(fn)
    def preconfigured_function(*fn_args, **fn_kwargs):
      if tf.executing_eagerly():
        return fn(*fn_args, **fn_kwargs)
      if wrapped[0] is None:
        wrapped[0] = function(*((fn,) + args), **kwargs)
      return wrapped[0](*fn_args, **fn_kwargs)  # pylint: disable=not-callable

    return preconfigured_function

  return maybe_wrap


def function(*args, **kwargs):
  """Wrapper for tf.function with TF Agents-specific customizations.

  Example:

  ```python
  @common.function()
  def my_eager_code(x, y):
    ...
  ```

  Args:
    *args: Args for tf.function.
    **kwargs: Keyword args for tf.function.

  Returns:
    A tf.function wrapper.
  """
  autograph = kwargs.pop('autograph', False)
  reduce_retracing = kwargs.pop('reduce_retracing', True)
  return tf.function(  # allow-tf-function
      *args,
      autograph=autograph,
      reduce_retracing=reduce_retracing,
      **kwargs)


def has_eager_been_enabled():
  """Returns true iff in TF2 or in TF1 with eager execution enabled."""
  with tf.init_scope():
    return tf.executing_eagerly()


def function_in_tf1(*args, **kwargs):
  """Wrapper that returns common.function if using TF1.

  This allows for code that assumes autodeps is available to be written once,
  in the same way, for both TF1 and TF2.

  Usage:

  ```python
  train = function_in_tf1()(agent.train)
  loss = train(experience)
  ```

  Args:
    *args: Arguments for common.function.
    **kwargs: Keyword arguments for common.function.

  Returns:
    A callable that wraps a function.
  """

  def maybe_wrap(fn):
    """Helper function."""
    # We're in TF1 mode and want to wrap in common.function to get autodeps.
    wrapped = [None]

    @functools.wraps(fn)
    def with_check_resource_vars(*fn_args, **fn_kwargs):
      """Helper function for calling common.function."""
      check_tf1_allowed()
      if has_eager_been_enabled():
        # We're either in eager mode or in tf.function mode (no in-between); so
        # autodep-like behavior is already expected of fn.
        return fn(*fn_args, **fn_kwargs)
      if not resource_variables_enabled():
        raise RuntimeError(MISSING_RESOURCE_VARIABLES_ERROR)
      if wrapped[0] is None:
        wrapped[0] = function(*((fn,) + args), **kwargs)
      return wrapped[0](*fn_args, **fn_kwargs)  # pylint: disable=not-callable

    return with_check_resource_vars

  return maybe_wrap


def create_variable(name,
                    initial_value=0,
                    shape=(),
                    dtype=tf.int64,
                    use_local_variable=False,
                    trainable=False,
                    initializer=None,
                    unique_name=True):
  """Create a variable."""
  check_tf1_allowed()
  if has_eager_been_enabled():
    if initializer is None:
      if shape:
        initial_value = tf.constant(initial_value, shape=shape, dtype=dtype)
      else:
        initial_value = tf.convert_to_tensor(initial_value, dtype=dtype)
    else:
      if callable(initializer):
        initial_value = lambda: initializer(shape, dtype)
      else:
        initial_value = initializer
    return tf.compat.v2.Variable(
        initial_value, trainable=trainable, dtype=dtype, name=name)
  collections = [tf.compat.v1.GraphKeys.GLOBAL_VARIABLES]
  if use_local_variable:
    collections = [tf.compat.v1.GraphKeys.LOCAL_VARIABLES]
  if initializer is None:
    initializer = tf.compat.v1.initializers.constant(initial_value, dtype=dtype)
    if shape is None:
      shape = tf.convert_to_tensor(initial_value).shape
  if unique_name:
    name = tf.compat.v1.get_default_graph().unique_name(name)
  return tf.compat.v1.get_variable(
      name=name,
      shape=shape,
      dtype=dtype,
      initializer=initializer,
      collections=collections,
      use_resource=True,
      trainable=trainable)


def soft_variables_update(source_variables,
                          target_variables,
                          tau=1.0,
                          tau_non_trainable=None,
                          sort_variables_by_name=False):
  """Performs a soft/hard update of variables from the source to the target.

  Note: **when using this function with TF DistributionStrategy**, the
  `strategy.extended.update` call (below) needs to be done in a cross-replica
  context, i.e. inside a merge_call. Please use the Periodically class above
  that provides this wrapper for you.

  For each variable v_t in target variables and its corresponding variable v_s
  in source variables, a soft update is:
  v_t = (1 - tau) * v_t + tau * v_s

  When tau is 1.0 (the default), then it does a hard update:
  v_t = v_s

  Args:
    source_variables: list of source variables.
    target_variables: list of target variables.
    tau: A float scalar in [0, 1]. When tau is 1.0 (the default), we do a hard
      update. This is used for trainable variables.
    tau_non_trainable: A float scalar in [0, 1] for non_trainable variables. If
      None, will copy from tau.
    sort_variables_by_name: A bool, when True would sort the variables by name
      before doing the update.

  Returns:
    An operation that updates target variables from source variables.

  Raises:
    ValueError: if `tau not in [0, 1]`.
    ValueError: if `len(source_variables) != len(target_variables)`.
    ValueError: "Method requires being in cross-replica context,
      use get_replica_context().merge_call()" if used inside replica context.
  """
  if tau < 0 or tau > 1:
    raise ValueError('Input `tau` should be in [0, 1].')
  if tau_non_trainable is None:
    tau_non_trainable = tau

  if tau_non_trainable < 0 or tau_non_trainable > 1:
    raise ValueError('Input `tau_non_trainable` should be in [0, 1].')

  updates = []

  op_name = 'soft_variables_update'
  if tau == 0.0 or not source_variables or not target_variables:
    return tf.no_op(name=op_name)
  if len(source_variables) != len(target_variables):
    raise ValueError(
        'Source and target variable lists have different lengths: '
        '{} vs. {}'.format(len(source_variables), len(target_variables)))
  if sort_variables_by_name:
    source_variables = sorted(source_variables, key=lambda x: x.name)
    target_variables = sorted(target_variables, key=lambda x: x.name)

  strategy = tf.distribute.get_strategy()

  for (v_s, v_t) in zip(source_variables, target_variables):
    v_t.shape.assert_is_compatible_with(v_s.shape)

    def update_fn(v1, v2):
      """Update variables."""
      # For not trainable variables do hard updates.
      # This helps stabilaze BatchNorm moving averagees TODO(b/144455039)
      if not v1.trainable:
        current_tau = tau_non_trainable
      else:
        current_tau = tau

      if current_tau == 1.0:
        return v1.assign(v2)
      else:
        return v1.assign((1 - current_tau) * v1 + current_tau * v2)

    # TODO(b/142508640): remove this when b/142802462 is fixed.
    # Workaround for b/142508640, only use extended.update for
    # MirroredVariable variables (which are trainable variables).
    # For other types of variables (i.e. SyncOnReadVariables, for example
    # batch norm stats) do a regular assign, which will cause a sync and
    # broadcast from replica 0, so will have slower performance but will be
    # correct and not cause a failure.
    if tf.distribute.has_strategy() and v_t.trainable:
      # Assignment happens independently on each replica,
      # see b/140690837 #46.
      update = strategy.extended.update(v_t, update_fn, args=(v_s,))
    else:
      update = update_fn(v_t, v_s)

    updates.append(update)
  return tf.group(*updates, name=op_name)


def join_scope(parent_scope, child_scope):
  """Joins a parent and child scope using `/`, checking for empty/none.

  Args:
    parent_scope: (string) parent/prefix scope.
    child_scope: (string) child/suffix scope.

  Returns:
    joined scope: (string) parent and child scopes joined by /.
  """
  if not parent_scope:
    return child_scope
  if not child_scope:
    return parent_scope
  return '/'.join([parent_scope, child_scope])


# TODO(b/138322868): Add an optional action_spec for validation.
def index_with_actions(q_values, actions, multi_dim_actions=False):
  """Index into q_values using actions.

  Note: this supports multiple outer dimensions (e.g. time, batch etc).

  Args:
    q_values: A float tensor of shape [outer_dim1, ... outer_dimK, action_dim1,
      ..., action_dimJ].
    actions: An int tensor of shape [outer_dim1, ... outer_dimK]    if
      multi_dim_actions=False [outer_dim1, ... outer_dimK, J] if
      multi_dim_actions=True I.e. in the multidimensional case,
      actions[outer_dim1, ... outer_dimK] is a vector [actions_1, ...,
      actions_J] where each element actions_j is an action in the range [0,
      num_actions_j). While in the single dimensional case, actions[outer_dim1,
      ... outer_dimK] is a scalar.
    multi_dim_actions: whether the actions are multidimensional.

  Returns:
    A [outer_dim1, ... outer_dimK] tensor of q_values for the given actions.

  Raises:
    ValueError: If actions have unknown rank.
  """
  if actions.shape.rank is None:
    raise ValueError('actions should have known rank.')
  batch_dims = actions.shape.rank
  if multi_dim_actions:
    # In the multidimensional case, the last dimension of actions indexes the
    # vector of actions for each batch, so exclude it from the batch dimensions.
    batch_dims -= 1

  outer_shape = tf.shape(input=actions)
  batch_indices = tf.meshgrid(
      *[tf.range(outer_shape[i]) for i in range(batch_dims)], indexing='ij')
  batch_indices = [tf.cast(tf.expand_dims(batch_index, -1), dtype=tf.int32)
                   for batch_index in batch_indices]
  if not multi_dim_actions:
    actions = tf.expand_dims(actions, -1)
  # Cast actions to tf.int32 in order to avoid a TypeError in tf.concat.
  actions = tf.cast(actions, dtype=tf.int32)
  action_indices = tf.concat(batch_indices + [actions], -1)
  return tf.gather_nd(q_values, action_indices)


def periodically(body, period, name='periodically'):
  """Periodically performs the tensorflow op in `body`.

  The body tensorflow op will be executed every `period` times the periodically
  op is executed. More specifically, with `n` the number of times the op has
  been executed, the body will be executed when `n` is a non zero positive
  multiple of `period` (i.e. there exist an integer `k > 0` such that
  `k * period == n`).

  If `period` is `None`, it will not perform any op and will return a
  `tf.no_op()`.

  If `period` is 1, it will just execute the body, and not create any counters
  or conditionals.

  Args:
    body: callable that returns the tensorflow op to be performed every time an
      internal counter is divisible by the period. The op must have no output
      (for example, a tf.group()).
    period: inverse frequency with which to perform the op.
    name: name of the variable_scope.

  Raises:
    TypeError: if body is not a callable.

  Returns:
    An op that periodically performs the specified op.
  """
  if tf.executing_eagerly():
    if isinstance(period, tf.Variable):
      return Periodically(body, period, name)
    return EagerPeriodically(body, period)
  else:
    return Periodically(body, period, name)()


class Periodically(tf.Module):
  """Periodically performs the ops defined in `body`."""

  def __init__(self, body, period, name='periodically'):
    """Periodically performs the ops defined in `body`.

    The body tensorflow op will be executed every `period` times the
    periodically op is executed. More specifically, with `n` the number of times
    the op has been executed, the body will be executed when `n` is a non zero
    positive multiple of `period` (i.e. there exist an integer `k > 0` such that
    `k * period == n`).

    If `period` is `None`, it will not perform any op and will return a
    `tf.no_op()`.

    If `period` is 1, it will just execute the body, and not create any counters
    or conditionals.

    Args:
      body: callable that returns the tensorflow op to be performed every time
        an internal counter is divisible by the period. The op must have no
        output (for example, a tf.group()).
      period: inverse frequency with which to perform the op. It can be a Tensor
        or a Variable.
      name: name of the object.

    Raises:
      TypeError: if body is not a callable.

    Returns:
      An op that periodically performs the specified op.
    """
    super(Periodically, self).__init__(name=name)
    if not callable(body):
      raise TypeError('body must be callable.')
    self._body = body
    self._period = period
    self._counter = create_variable(self.name + '/counter', 0)

  def __call__(self):

    def call(strategy=None):
      del strategy  # unused
      if self._period is None:
        return tf.no_op()
      if self._period == 1:
        return self._body()
      period = tf.cast(self._period, self._counter.dtype)
      remainder = tf.math.mod(self._counter.assign_add(1), period)
      return tf.cond(
          pred=tf.equal(remainder, 0), true_fn=self._body, false_fn=tf.no_op)

    # TODO(b/129083817) add an explicit unit test to ensure correct behavior
    ctx = tf.distribute.get_replica_context()
    if ctx:
      return tf.distribute.get_replica_context().merge_call(call)
    else:
      return call()


class EagerPeriodically(object):
  """EagerPeriodically performs the ops defined in `body`.

  Only works in Eager mode.
  """

  def __init__(self, body, period):
    """EagerPeriodically performs the ops defined in `body`.

    Args:
      body: callable that returns the tensorflow op to be performed every time
        an internal counter is divisible by the period. The op must have no
        output (for example, a tf.group()).
      period: inverse frequency with which to perform the op. Must be a simple
        python int/long.

    Raises:
      TypeError: if body is not a callable.

    Returns:
      An op that periodically performs the specified op.
    """
    if not callable(body):
      raise TypeError('body must be callable.')
    self._body = body
    self._period = period
    self._counter = 0

  def __call__(self):
    if self._period is None:
      return tf.no_op()
    if self._period == 1:
      return self._body()
    self._counter += 1
    if self._counter % self._period == 0:
      self._body()


def clip_to_spec(value, spec):
  """Clips value to a given bounded tensor spec.

  Args:
    value: (tensor) value to be clipped.
    spec: (BoundedTensorSpec) spec containing min. and max. values for clipping.

  Returns:
    clipped_value: (tensor) `value` clipped to be compatible with `spec`.
  """
  return tf.clip_by_value(value, spec.minimum, spec.maximum)


def spec_means_and_magnitudes(action_spec):
  """Get the center and magnitude of the ranges in action spec."""
  action_means = tf.nest.map_structure(
      lambda spec: (spec.maximum + spec.minimum) / 2.0, action_spec)
  action_magnitudes = tf.nest.map_structure(
      lambda spec: (spec.maximum - spec.minimum) / 2.0, action_spec)
  return np.array(
      action_means, dtype=np.float32), np.array(
          action_magnitudes, dtype=np.float32)


def scale_to_spec(tensor, spec):
  """Shapes and scales a batch into the given spec bounds.

  Args:
    tensor: A [batch x n] tensor with values in the range of [-1, 1].
    spec: (BoundedTensorSpec) to use for scaling the action.

  Returns:
    A batch scaled the given spec bounds.
  """
  tensor = tf.reshape(tensor, [-1] + spec.shape.as_list())

  # Scale the tensor.
  means, magnitudes = spec_means_and_magnitudes(spec)
  tensor = means + magnitudes * tensor

  # Set type.
  return tf.cast(tensor, spec.dtype)


def ornstein_uhlenbeck_process(initial_value,
                               damping=0.15,
                               stddev=0.2,
                               seed=None,
                               scope='ornstein_uhlenbeck_noise'):
  """An op for generating noise from a zero-mean Ornstein-Uhlenbeck process.

  The Ornstein-Uhlenbeck process is a process that generates temporally
  correlated noise via a random walk with damping. This process describes
  the velocity of a particle undergoing brownian motion in the presence of
  friction. This can be useful for exploration in continuous action environments
  with momentum.

  The temporal update equation is:
  `x_next = (1 - damping) * x + N(0, std_dev)`

  Args:
    initial_value: Initial value of the process.
    damping: The rate at which the noise trajectory is damped towards the mean.
      We must have 0 <= damping <= 1, where a value of 0 gives an undamped
      random walk and a value of 1 gives uncorrelated Gaussian noise. Hence in
      most applications a small non-zero value is appropriate.
    stddev: Standard deviation of the Gaussian component.
    seed: Seed for random number generation.
    scope: Scope of the variables.

  Returns:
    An op that generates noise.
  """
  if tf.executing_eagerly():
    return OUProcess(initial_value, damping, stddev, seed, scope)
  else:
    return OUProcess(initial_value, damping, stddev, seed, scope)()


class OUProcess(tf.Module):
  """A zero-mean Ornstein-Uhlenbeck process."""

  def __init__(self,
               initial_value,
               damping=0.15,
               stddev=0.2,
               seed=None,
               scope='ornstein_uhlenbeck_noise'):
    """A Class for generating noise from a zero-mean Ornstein-Uhlenbeck process.

    The Ornstein-Uhlenbeck process is a process that generates temporally
    correlated noise via a random walk with damping. This process describes
    the velocity of a particle undergoing brownian motion in the presence of
    friction. This can be useful for exploration in continuous action
    environments with momentum.

    The temporal update equation is:
    `x_next = (1 - damping) * x + N(0, std_dev)`

    Args:
      initial_value: Initial value of the process.
      damping: The rate at which the noise trajectory is damped towards the
        mean. We must have 0 <= damping <= 1, where a value of 0 gives an
        undamped random walk and a value of 1 gives uncorrelated Gaussian noise.
        Hence in most applications a small non-zero value is appropriate.
      stddev: Standard deviation of the Gaussian component.
      seed: Seed for random number generation.
      scope: Scope of the variables.
    """
    super(OUProcess, self).__init__()
    self._damping = damping
    self._stddev = stddev
    self._seed = seed
    with tf.name_scope(scope):
      self._x = tf.compat.v2.Variable(
          initial_value=initial_value, trainable=False)

  def __call__(self):
    noise = tf.random.normal(
        shape=self._x.shape,
        stddev=self._stddev,
        dtype=self._x.dtype,
        seed=self._seed)
    return self._x.assign((1. - self._damping) * self._x + noise)


def log_probability(distributions, actions, action_spec):
  """Computes log probability of actions given distribution.

  Args:
    distributions: A possibly batched tuple of distributions.
    actions: A possibly batched action tuple.
    action_spec: A nested tuple representing the action spec.

  Returns:
    A Tensor representing the log probability of each action in the batch.
  """
  outer_rank = nest_utils.get_outer_rank(actions, action_spec)

  def _compute_log_prob(single_distribution, single_action):
    # sum log-probs over everything but the batch
    single_log_prob = single_distribution.log_prob(single_action)
    rank = single_log_prob.shape.rank
    reduce_dims = list(range(outer_rank, rank))
    return tf.reduce_sum(
        input_tensor=single_log_prob,
        axis=reduce_dims)

  nest_utils.assert_same_structure(distributions, actions)
  log_probs = [
      _compute_log_prob(dist, action)
      for (dist, action
          ) in zip(tf.nest.flatten(distributions), tf.nest.flatten(actions))
  ]

  # sum log-probs over action tuple
  total_log_probs = tf.add_n(log_probs)

  return total_log_probs


# TODO(ofirnachum): Move to distribution utils.
def entropy(distributions, action_spec, outer_rank=None):
  """Computes total entropy of distribution.

  Args:
    distributions: A possibly batched tuple of distributions.
    action_spec: A nested tuple representing the action spec.
    outer_rank: Optional outer rank of the distributions. If not provided use
      distribution.mode() to compute it.

  Returns:
    A Tensor representing the entropy of each distribution in the batch.
    Assumes actions are independent, so that marginal entropies of each action
    may be summed.
  """
  if outer_rank is None:
    nested_modes = tf.nest.map_structure(lambda d: d.mode(), distributions)
    outer_rank = nest_utils.get_outer_rank(nested_modes, action_spec)

  def _compute_entropy(single_distribution):
    try:
      entropies = single_distribution.entropy()
      # Sum entropies over everything but the batch.
      rank = entropies.shape.rank
      reduce_dims = list(range(outer_rank, rank))
      return tf.reduce_sum(input_tensor=entropies, axis=reduce_dims)
    except NotImplementedError:
      return None

  entropies = []
  for dist in tf.nest.flatten(distributions):
    entropy_dist = _compute_entropy(dist)
    if entropy_dist is not None:
      entropies.append(entropy_dist)

  # Sum entropies over action tuple.
  if not entropies:
    return None

  return tf.add_n(entropies)


def discounted_future_sum(values, gamma, num_steps):
  """Discounted future sum of batch-major values.

  Args:
    values: A Tensor of shape [batch_size, total_steps] and dtype float32.
    gamma: A float discount value.
    num_steps: A positive integer number of future steps to sum.

  Returns:
    A Tensor of shape [batch_size, total_steps], where each entry `(i, j)` is
      the result of summing the entries of values starting from
      `gamma^0 * values[i, j]` to
      `gamma^(num_steps - 1) * values[i, j + num_steps - 1]`,
      with zeros padded to values.

      For example, values=[5, 6, 7], gamma=0.9, will result in sequence:
      ```python
      [(5 * 0.9^0 + 6 * 0.9^1 + 7 * 0.9^2), (6 * 0.9^0 + 7 * 0.9^1), 7 * 0.9^0]
      ```

  Raises:
    ValueError: If values is not of rank 2.
  """
  if values.get_shape().rank != 2:
    raise ValueError('Input must be rank 2 tensor.  Got %d.' %
                     values.get_shape().rank)

  (batch_size, total_steps) = values.get_shape().as_list()

  num_steps = tf.minimum(num_steps, total_steps)
  discount_filter = tf.reshape(gamma**tf.cast(tf.range(num_steps), tf.float32),
                               [-1, 1, 1])
  padded_values = tf.concat([values, tf.zeros([batch_size, num_steps - 1])], 1)

  convolved_values = tf.squeeze(
      tf.nn.conv1d(
          input=tf.expand_dims(padded_values, -1),
          filters=discount_filter,
          stride=1,
          padding='VALID'), -1)

  return convolved_values


def discounted_future_sum_masked(values, gamma, num_steps, episode_lengths):
  """Discounted future sum of batch-major values.

  Args:
    values: A Tensor of shape [batch_size, total_steps] and dtype float32.
    gamma: A float discount value.
    num_steps: A positive integer number of future steps to sum.
    episode_lengths: A vector shape [batch_size] with num_steps per episode.

  Returns:
    A Tensor of shape [batch_size, total_steps], where each entry is the
      discounted sum as in discounted_future_sum, except with values after
      the end of episode_lengths masked to 0.

  Raises:
    ValueError: If values is not of rank 2, or if total_steps is not defined.
  """
  if values.shape.rank != 2:
    raise ValueError('Input must be a rank 2 tensor.  Got %d.' % values.shape)

  total_steps = tf.compat.dimension_value(values.shape[1])
  if total_steps is None:
    raise ValueError('total_steps dimension in input '
                     'values[batch_size, total_steps] must be fully defined.')

  episode_mask = tf.cast(
      tf.sequence_mask(episode_lengths, total_steps), tf.float32)
  values *= episode_mask
  return discounted_future_sum(values, gamma, num_steps)


def shift_values(values, gamma, num_steps, final_values=None):
  """Shifts batch-major values in time by some amount.

  Args:
    values: A Tensor of shape [batch_size, total_steps] and dtype float32.
    gamma: A float discount value.
    num_steps: A nonnegative integer amount to shift values by.
    final_values: A float32 Tensor of shape [batch_size] corresponding to the
      values at step num_steps + 1.  Defaults to None (all zeros).

  Returns:
    A Tensor of shape [batch_size, total_steps], where each entry (i, j) is
    gamma^num_steps * values[i, j + num_steps] if j + num_steps < total_steps;
    gamma^(total_steps - j) * final_values[i] otherwise.

  Raises:
    ValueError: If values is not of rank 2.
  """
  if values.get_shape().rank != 2:
    raise ValueError('Input must be rank 2 tensor.  Got %d.' %
                     values.get_shape().rank)

  (batch_size, total_steps) = values.get_shape().as_list()
  num_steps = tf.minimum(num_steps, total_steps)

  if final_values is None:
    final_values = tf.zeros([batch_size])

  padding_exponent = tf.expand_dims(
      tf.cast(tf.range(num_steps, 0, -1), tf.float32), 0)
  final_pad = tf.expand_dims(final_values, 1) * gamma**padding_exponent
  return tf.concat([
      gamma**tf.cast(num_steps, tf.float32) * values[:, num_steps:], final_pad
  ], 1)


def get_episode_mask(time_steps):
  """Create a mask that is 0.0 for all final steps, 1.0 elsewhere.

  Args:
    time_steps: A TimeStep namedtuple representing a batch of steps.

  Returns:
    A float32 Tensor with 0s where step_type == LAST and 1s otherwise.
  """
  episode_mask = tf.cast(
      tf.not_equal(time_steps.step_type, ts.StepType.LAST), tf.float32)
  return episode_mask


def get_contiguous_sub_episodes(next_time_steps_discount):
  """Computes mask on sub-episodes which includes only contiguous components.

  Args:
    next_time_steps_discount: Tensor of shape [batch_size, total_steps]
      corresponding to environment discounts on next time steps (i.e.
      next_time_steps.discount).

  Returns:
    A float Tensor of shape [batch_size, total_steps] specifying mask including
      only contiguous components. Each row will be of the form
      [1.0] * a + [0.0] * b, where a >= 1 and b >= 0, and in which the initial
      sequence of ones corresponds to a contiguous sub-episode.
  """
  episode_end = tf.equal(next_time_steps_discount,
                         tf.constant(0, dtype=next_time_steps_discount.dtype))
  mask = tf.math.cumprod(
      1.0 - tf.cast(episode_end, tf.float32), axis=1, exclusive=True)
  return mask


def convert_q_logits_to_values(logits, support):
  """Converts a set of Q-value logits into Q-values using the provided support.

  Args:
    logits: A Tensor representing the Q-value logits.
    support: The support of the underlying distribution.

  Returns:
    A Tensor containing the expected Q-values.
  """
  probabilities = tf.nn.softmax(logits)
  return tf.reduce_sum(input_tensor=support * probabilities, axis=-1)


def generate_tensor_summaries(tag, tensor, step):
  """Generates various summaries of `tensor` such as histogram, max, min, etc.

  Args:
    tag: A namescope tag for the summaries.
    tensor: The tensor to generate summaries of.
    step: Variable to use for summaries.
  """
  with tf.name_scope(tag):
    tf.compat.v2.summary.histogram(name='histogram', data=tensor, step=step)
    tf.compat.v2.summary.scalar(
        name='mean', data=tf.reduce_mean(input_tensor=tensor), step=step)
    tf.compat.v2.summary.scalar(
        name='mean_abs',
        data=tf.reduce_mean(input_tensor=tf.abs(tensor)),
        step=step)
    tf.compat.v2.summary.scalar(
        name='max', data=tf.reduce_max(input_tensor=tensor), step=step)
    tf.compat.v2.summary.scalar(
        name='min', data=tf.reduce_min(input_tensor=tensor), step=step)
    tf.compat.v2.summary.scalar(
        name='std', data=tf.math.reduce_std(input_tensor=tensor), step=step)


def summarize_tensor_dict(tensor_dict: Dict[Text, types.Tensor],
                          step: Optional[types.Tensor]):
  """Generates summaries of all tensors in `tensor_dict`.

  Args:
    tensor_dict: A dictionary {name, tensor} to summarize.
    step: The global step
  """
  for tag in tensor_dict:
    generate_tensor_summaries(tag, tensor_dict[tag], step)


def compute_returns(rewards: types.Tensor,
                    discounts: types.Tensor,
                    time_major: bool = False):
  """Compute the return from each index in an episode.

  Args:
    rewards: Tensor `[T]`, `[B, T]`, `[T, B]` of per-timestep reward.
    discounts: Tensor `[T]`, `[B, T]`, `[T, B]` of per-timestep discount factor.
      Should be `0`. for final step of each episode.
    time_major: Bool, when batched inputs setting it to `True`, inputs are
      expected to be time-major: `[T, B]` otherwise, batch-major: `[B, T]`.

  Returns:
    Tensor of per-timestep cumulative returns.
  """
  rewards.shape.assert_is_compatible_with(discounts.shape)
  if (not rewards.shape.is_fully_defined() or
      not discounts.shape.is_fully_defined()):
    tf.debugging.assert_equal(tf.shape(input=rewards),
                              tf.shape(input=discounts))

  def discounted_accumulate_rewards(next_step_return, reward_and_discount):
    reward, discount = reward_and_discount
    return next_step_return * discount + reward

  # Support batched rewards and discount via transpose.
  if rewards.shape.rank > 1 and not time_major:
    rewards = tf.transpose(rewards, perm=[1, 0])
    discounts = tf.transpose(discounts, perm=[1, 0])
  # Cumulatively sum discounted reward R_t.
  #   R_t = r_t + discount * (r_t+1 + discount * (r_t+2 * discount( ...
  # As discount is 0 for terminal states, ends of episode will not include
  #   reward from subsequent timesteps.
  returns = tf.scan(
      discounted_accumulate_rewards, [rewards, discounts],
      initializer=tf.zeros_like(rewards[0]),
      reverse=True)
  # Reverse transpose if needed.
  if returns.shape.rank > 1 and not time_major:
    returns = tf.transpose(returns, perm=[1, 0])
  return returns


def initialize_uninitialized_variables(session, var_list=None):
  """Initialize any pending variables that are uninitialized."""
  if var_list is None:
    var_list = tf.compat.v1.global_variables() + tf.compat.v1.local_variables()
  is_initialized = session.run(
      [tf.compat.v1.is_variable_initialized(v) for v in var_list])
  uninitialized_vars = []
  for flag, v in zip(is_initialized, var_list):
    if not flag:
      uninitialized_vars.append(v)
  if uninitialized_vars:
    logging.info('uninitialized_vars: %s',
                 ', '.join([str(x) for x in uninitialized_vars]))
    session.run(tf.compat.v1.variables_initializer(uninitialized_vars))


class Checkpointer(object):
  """Checkpoints training state, policy state, and replay_buffer state."""

  def __init__(self, ckpt_dir, max_to_keep=20, **kwargs):
    """A class for making checkpoints.

    If ckpt_dir doesn't exists it creates it.

    Args:
      ckpt_dir: The directory to save checkpoints.
      max_to_keep: Maximum number of checkpoints to keep (if greater than the
        max are saved, the oldest checkpoints are deleted).
      **kwargs: Items to include in the checkpoint.
    """
    self._checkpoint = tf.train.Checkpoint(**kwargs)

    if not tf.io.gfile.exists(ckpt_dir):
      tf.io.gfile.makedirs(ckpt_dir)

    self._manager = tf.train.CheckpointManager(
        self._checkpoint, directory=ckpt_dir, max_to_keep=max_to_keep)

    if self._manager.latest_checkpoint is not None:
      logging.info('Checkpoint available: %s', self._manager.latest_checkpoint)
      self._checkpoint_exists = True
    else:
      logging.info('No checkpoint available at %s', ckpt_dir)
      self._checkpoint_exists = False
    self._load_status = self._checkpoint.restore(
        self._manager.latest_checkpoint)

  @property
  def checkpoint_exists(self):
    return self._checkpoint_exists

  @property
  def manager(self):
    """Returns the underlying tf.train.CheckpointManager."""
    return self._manager

  def initialize_or_restore(self, session=None):
    """Initialize or restore graph (based on checkpoint if exists)."""
    self._load_status.initialize_or_restore(session)
    return self._load_status

  def save(self, global_step: tf.Tensor,
           options: tf.train.CheckpointOptions = None):
    """Save state to checkpoint."""
    saved_checkpoint = self._manager.save(
        checkpoint_number=global_step, options=options)
    self._checkpoint_exists = True
    logging.info('%s', 'Saved checkpoint: {}'.format(saved_checkpoint))


def replicate(tensor, outer_shape):
  """Replicates a tensor so as to match the given outer shape.

  Example:
  - t = [[1, 2, 3], [4, 5, 6]] (shape = [2, 3])
  - outer_shape = [2, 1]
  The shape of the resulting tensor is: [2, 1, 2, 3]
  and its content is: [[t], [t]]

  Args:
    tensor: A tf.Tensor.
    outer_shape: Outer shape given as a 1D tensor of type list, numpy or
      tf.Tensor.

  Returns:
    The replicated tensor.

  Raises:
    ValueError: when the outer shape is incorrect.
  """
  outer_shape = tf.convert_to_tensor(value=outer_shape)
  if len(outer_shape.shape) != 1:
    raise ValueError('The outer shape must be a 1D tensor')
  outer_ndims = int(outer_shape.shape[0])
  tensor_ndims = len(tensor.shape)

  # No need to replicate anything if there is no outer dim to add.
  if outer_ndims == 0:
    return tensor

  # Calculate target shape of replicated tensor
  target_shape = tf.concat([outer_shape, tf.shape(input=tensor)], axis=0)

  # tf.tile expects `tensor` to be at least 1D
  if tensor_ndims == 0:
    tensor = tensor[None]

  # Replicate tensor "t" along the 1st dimension.
  tiled_tensor = tf.tile(tensor, [tf.reduce_prod(input_tensor=outer_shape)] +
                         [1] * (tensor_ndims - 1))

  # Reshape to match outer_shape.
  return tf.reshape(tiled_tensor, target_shape)


def assert_members_are_not_overridden(base_cls,
                                      instance,
                                      allowlist=(),
                                      denylist=()):
  """Asserts public members of `base_cls` are not overridden in `instance`.

  If both `allowlist` and `denylist` are empty, no public member of
  `base_cls` can be overridden. If a `allowlist` is provided, only public
  members in `allowlist` can be overridden. If a `denylist` is provided,
  all public members except those in `denylist` can be overridden. Both
  `allowlist` and `denylist` cannot be provided at the same, if so a
  ValueError will be raised.

  Args:
    base_cls: A Base class.
    instance: An instance of a subclass of `base_cls`.
    allowlist: Optional list of `base_cls` members that can be overridden.
    denylist: Optional list of `base_cls` members that cannot be overridden.

  Raises:
    ValueError if both allowlist and denylist are provided.
  """

  if denylist and allowlist:
    raise ValueError('Both `denylist` and `allowlist` cannot be provided.')

  instance_type = type(instance)
  subclass_members = set(instance_type.__dict__.keys())
  public_members = set(
      [m for m in base_cls.__dict__.keys() if not m.startswith('_')])
  common_members = public_members & subclass_members

  if allowlist:
    common_members = common_members - set(allowlist)
  elif denylist:
    common_members = common_members & set(denylist)

  overridden_members = [
      m for m in common_members
      if base_cls.__dict__[m] != instance_type.__dict__[m]
  ]
  if overridden_members:
    raise ValueError(
        'Subclasses of {} cannot override most of its base members, but '
        '{} overrides: {}'.format(base_cls, instance_type, overridden_members))


def element_wise_squared_loss(x, y):
  return tf.compat.v1.losses.mean_squared_error(
      x, y, reduction=tf.compat.v1.losses.Reduction.NONE)


def element_wise_huber_loss(x, y):
  return tf.compat.v1.losses.huber_loss(
      x, y, reduction=tf.compat.v1.losses.Reduction.NONE)


def transpose_batch_time(x):
  """Transposes the batch and time dimensions of a Tensor.

  If the input tensor has rank < 2 it returns the original tensor. Retains as
  much of the static shape information as possible.

  Args:
    x: A Tensor.

  Returns:
    x transposed along the first two dimensions.
  """
  x_static_shape = x.get_shape()
  if x_static_shape.rank is not None and x_static_shape.rank < 2:
    return x

  x_rank = tf.rank(x)
  x_t = tf.transpose(a=x, perm=tf.concat(([1, 0], tf.range(2, x_rank)), axis=0))
  x_t.set_shape(
      tf.TensorShape(
          [x_static_shape.dims[1].value,
           x_static_shape.dims[0].value]).concatenate(x_static_shape[2:]))
  return x_t


def save_spec(spec, file_path):
  """Saves the given spec nest as a StructProto.

  **Note**: Currently this will convert BoundedTensorSpecs into regular
    TensorSpecs.

  Args:
    spec: A nested structure of TensorSpecs.
    file_path: Path to save the encoded spec to.
  """
  spec = tensor_spec.from_spec(spec)
  spec_proto = nested_structure_coder.encode_structure(spec)

  dir_path = os.path.dirname(file_path)
  if not tf.io.gfile.exists(dir_path):
    tf.io.gfile.makedirs(dir_path)

  with tf.compat.v2.io.gfile.GFile(file_path, 'wb') as gfile:
    gfile.write(spec_proto.SerializeToString())


def load_spec(file_path):
  """Loads a data spec from a file.

  **Note**: Types for Named tuple classes will not match. Users need to convert
    to these manually:

    # Convert from:
    # 'tensorflow.python.saved_model.nested_structure_coder.Trajectory'
    # to proper TrajectorySpec.
    # trajectory_spec = trajectory.Trajectory(*spec)

  Args:
    file_path: Path to the saved data spec.
  Returns:
    A nested structure of TensorSpecs.
  """
  with tf.compat.v2.io.gfile.GFile(file_path, 'rb') as gfile:
    signature_proto = struct_pb2.StructuredValue.FromString(gfile.read())

  return nested_structure_coder.decode_proto(signature_proto)


def extract_shared_variables(variables_1, variables_2):
  """Separates shared variables from the given collections.

  Args:
    variables_1: An iterable of Variables
    variables_2: An iterable of Variables

  Returns:
    A Tuple of ObjectIdentitySets described by the set operations

    ```
    (variables_1 - variables_2,
     variables_2 - variables_1,
     variables_1 & variables_2)
    ```
  """
  var_refs1 = object_identity.ObjectIdentitySet(variables_1)
  var_refs2 = object_identity.ObjectIdentitySet(variables_2)

  shared_vars = var_refs1.intersection(var_refs2)
  return (var_refs1.difference(shared_vars), var_refs2.difference(shared_vars),
          shared_vars)


def check_no_shared_variables(network_1, network_2):
  """Checks that there are no shared trainable variables in the two networks.

  Args:
    network_1: A network.Network.
    network_2: A network.Network.

  Raises:
    ValueError: if there are any common trainable variables.
    ValueError: if one of the networks has not yet been built
      (e.g. user must call `create_variables`).
  """
  variables_1 = object_identity.ObjectIdentitySet(network_1.trainable_variables)
  variables_2 = object_identity.ObjectIdentitySet(network_2.trainable_variables)
  shared_variables = variables_1 & variables_2
  if shared_variables:
    raise ValueError(
        'After making a copy of network \'{}\' to create a target '
        'network \'{}\', the target network shares weights with '
        'the original network.  This is not allowed.  If '
        'you want explicitly share weights with the target network, or '
        'if your input network shares weights with others, please '
        'provide a target network which explicitly, selectively, shares '
        'layers/weights with the input network.  If you are not intending to '
        'share weights make sure all the weights are created inside the Network'
        ' since a copy will be created by creating a new Network with the same '
        'args but a new name. Shared variables found: '
        '\'{}\'.'.format(
            network_1.name, network_2.name,
            [x.name for x in shared_variables]))


def check_matching_networks(network_1, network_2):
  """Check that two networks have matching input specs and variables.

  Args:
    network_1: A network.Network.
    network_2: A network.Network.

  Raises:
    ValueError: if the networks differ in input_spec, variables (number, dtype,
      or shape).
    ValueError: if either of the networks has not been built yet
      (e.g. user must call `create_variables`).
  """
  if network_1.input_tensor_spec != network_2.input_tensor_spec:
    raise ValueError('Input tensor specs of network and target network '
                     'do not match: {} vs. {}.'.format(
                         network_1.input_tensor_spec,
                         network_2.input_tensor_spec))
  if len(network_1.variables) != len(network_2.variables):
    raise ValueError(
        'Variables lengths do not match between Q network and target network: '
        '{} vs. {}'.format(network_1.variables, network_2.variables))
  for v1, v2 in zip(network_1.variables, network_2.variables):
    if v1.dtype != v2.dtype or v1.shape != v2.shape:
      raise ValueError(
          'Variable dtypes or shapes do not match: {} vs. {}'.format(v1, v2))


def maybe_copy_target_network_with_checks(network, target_network=None,
                                          name=None,
                                          input_spec=None):
  """Copies the network into target if None and checks for shared variables."""
  if target_network is None:
    target_network = network.copy(name=name)
    target_network.create_variables(input_spec)
  # Copy may have been shallow, and variables may inadvertently be shared
  # between the target and the original networks. This would be an unusual
  # setup, so we throw an error to protect users from accidentally doing so.
  # If you explicitly want this to be enabled, please open a feature request
  # with the team.
  check_no_shared_variables(network, target_network)
  check_matching_networks(network, target_network)
  return target_network


AggregatedLosses = cs.namedtuple(
    'AggregatedLosses',
    ['total_loss',  # Total loss = weighted + regularization
     'weighted',  # Weighted sum of per_example_loss by sample_weight.
     'regularization',  # Total of regularization losses.
    ])


def aggregate_losses(per_example_loss=None,
                     sample_weight=None,
                     global_batch_size=None,
                     regularization_loss=None):
  """Aggregates and scales per example loss and regularization losses.

  If `global_batch_size` is given it would be used for scaling, otherwise it
  would use the batch_dim of per_example_loss and number of replicas.

  Args:
    per_example_loss: Per-example loss [B] or [B, T, ...].
    sample_weight: Optional weighting for each example, Tensor shaped [B] or
      [B, T, ...], or a scalar float.
    global_batch_size: Optional global batch size value. Defaults to (size of
    first dimension of `losses`) * (number of replicas).
    regularization_loss: Regularization loss.

  Returns:
    An AggregatedLosses named tuple with scalar losses to optimize.
  """
  total_loss, weighted_loss, reg_loss = None, None, None
  if sample_weight is not None and not isinstance(sample_weight, tf.Tensor):
    sample_weight = tf.convert_to_tensor(sample_weight, dtype=tf.float32)

  # Compute loss that is scaled by global batch size.
  if per_example_loss is not None:
    loss_rank = per_example_loss.shape.rank
    if sample_weight is not None:
      weight_rank = sample_weight.shape.rank
      # Expand `sample_weight` to be broadcastable to the shape of
      # `per_example_loss`, to ensure that multiplication works properly.
      if weight_rank > 0 and loss_rank > weight_rank:
        for dim in range(weight_rank, loss_rank):
          sample_weight = tf.expand_dims(sample_weight, dim)
      # Sometimes we have an episode boundary or similar, and at this location
      # the loss is nonsensical (i.e., inf or nan); and sample_weight is zero.
      # In this case, we should respect the zero sample_weight and ignore the
      # frame.
      per_example_loss = tf.math.multiply_no_nan(
          per_example_loss, sample_weight)

    if loss_rank is not None and loss_rank == 0:
      err_msg = (
          'Need to use a loss function that computes losses per sample, ex: '
          'replace losses.mean_squared_error with tf.math.squared_difference. '
          'Invalid value passed for `per_example_loss`. Expected a tensor '
          'tensor with at least rank 1, received: {}'.format(per_example_loss))
      if tf.distribute.has_strategy():
        raise ValueError(err_msg)
      else:
        logging.warning(err_msg)
        # Add extra dimension to prevent error in compute_average_loss.
        per_example_loss = tf.expand_dims(per_example_loss, 0)
    elif loss_rank > 1:
      # If per_example_loss is shaped [B, T, ...], we need to compute the mean
      # across the extra dimensions, ex. time, as well.
      per_example_loss = tf.reduce_mean(per_example_loss, range(1, loss_rank))

    global_batch_size = global_batch_size and tf.cast(global_batch_size,
                                                      per_example_loss.dtype)
    weighted_loss = tf.nn.compute_average_loss(
        per_example_loss,
        global_batch_size=global_batch_size)
    total_loss = weighted_loss
  # Add scaled regularization losses.
  if regularization_loss is not None:
    reg_loss = tf.nn.scale_regularization_loss(regularization_loss)
    if total_loss is None:
      total_loss = reg_loss
    else:
      total_loss += reg_loss
  return AggregatedLosses(total_loss, weighted_loss, reg_loss)


def summarize_scalar_dict(name_data, step, name_scope='Losses/'):
  if name_data:
    with tf.name_scope(name_scope):
      for name, data in name_data.items():
        if data is not None:
          tf.compat.v2.summary.scalar(
              name=name, data=data, step=step)


@contextlib.contextmanager
def soft_device_placement():
  """Context manager for soft device placement, allowing summaries on CPU.

  Eager and graph contexts have different default device placements. See
  b/148408921 for details. This context manager should be used whenever using
  summary writers contexts to make sure summaries work when executing on TPUs.

  Yields:
    Sets `tf.config.set_soft_device_placement(True)` within the context
  """
  original_setting = tf.config.get_soft_device_placement()
  try:
    tf.config.set_soft_device_placement(True)
    yield
  finally:
    tf.config.set_soft_device_placement(original_setting)


def deduped_network_variables(network, *args):
  """Returns a list of variables in net1 that are not in any other nets.

  Args:
    network: A Keras network.
    *args: other networks to check for duplicate variables.
  """
  other_vars = object_identity.ObjectIdentitySet(
      [v for n in args for v in n.variables])  # pylint:disable=g-complex-comprehension
  return [v for v in network.variables if v not in other_vars]


def safe_has_state(state):
  """Safely checks `state not in (None, (), [])`."""
  # TODO(b/158804957): tf.function changes "s in ((),)" to a tensor bool expr.
  # pylint: disable=literal-comparison
  return state is not None and state is not () and state is not []
  # pylint: enable=literal-comparison