redis_table_ops.py

# Copyright 2020 The TensorFlow Authors. All Rights Reserved.
#
# 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
#
#     http://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.
# ==============================================================================
"""Redis Lookup operations."""
# pylint: disable=g-bad-name
from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

import fcntl
import functools
import json
import os
from re import T
import warnings

from tensorflow.python.eager import context
from tensorflow.python.framework import dtypes
from tensorflow.python.framework import ops
from tensorflow.python.ops import array_ops
from tensorflow.python.ops.lookup_ops import LookupInterface
from tensorflow.python.training.saver import BaseSaverBuilder

from tensorflow_recommenders_addons.utils.resource_loader import LazySO
from tensorflow_recommenders_addons.utils.resource_loader import prefix_op_name

redis_table_ops = LazySO("dynamic_embedding/core/_redis_table_ops.so").ops


class RedisTable(LookupInterface):
  """
    A generic mutable hash table implementation.

    Data can be inserted by calling the insert method and removed by calling the
    remove method. It does not support initialization via the init method.

    Example usage:

    ```python
    table = tfra.dynamic_embedding.RedisTable(key_dtype=tf.string,
                                                   value_dtype=tf.int64,
                                                   default_value=-1)
    sess.run(table.insert(keys, values))
    out = table.lookup(query_keys)
    print(out.eval())
    ```
  """

  default_redis_params = {
      "redis_connection_mode":
          1,  # ClusterMode = 0, SentinelMode = 1, StandaloneMode = 2
      "redis_master_name": "master",
      # connection_options
      "redis_host_ip": ["127.0.0.1"],
      "redis_host_port": [6379],
      "redis_user": "default",
      "redis_password": "",
      "redis_db": 0,
      "redis_read_access_slave":
          False,  # set True in infer or train mode if you like
      "redis_connect_keep_alive": False,  # keep TCP alive
      "redis_connect_timeout": 1000,  # milliseconds
      "redis_socket_timeout": 1000,  # milliseconds
      # connection_pool_options
      "redis_conn_pool_size": 20,
      "redis_wait_timeout": 100000000,  # milliseconds
      "redis_connection_lifetime": 100,  # minutes
      # sentinel_connection_options
      "redis_sentinel_user": "default",
      "redis_sentinel_password": "",
      "redis_sentinel_connect_timeout": 1000,  # milliseconds
      "redis_sentinel_socket_timeout": 1000,  # milliseconds
      # Below there is user-defined parameters in this custom op, not Redis setting parameters
      "storage_slice_import":
          -1,  # If storage_slice_import is not equal to storage_slice, rehash will happen. Equaling -1 means same as storage_slice.
      "storage_slice":
          1,  # For deciding bucket number, which usually is how many Redis instance may be used in the trainning.
      "using_hash_storage_slice":
          False,  # If True, IDs will be calculated hash(CRC32) value and then MOD to decide which bucket number they belong to. If False, only calculate the remainder.
      "keys_sending_size":
          1024,  # Determines how many keys to send at a time for performance tuning
      "using_md5_prefix_name": False,  # 1=true, 0=false
      "redis_hash_tags_hypodispersion":
          False,  # distribution of storag_slice will be hypodispersion in 16354 regardless cluster slot, but still depends on redis_hash_tags_import/runtime if they aren't empty.
      "model_tag_import":
          "test",  #  model_tag_import for version and any other information from last time.
      "redis_hash_tags_import": [
      ],  # Deciding hash tag for every bucket from last time, Note that the hash tag must be wrapped in curly braces {}.
      "model_tag_runtime":
          "test",  #  model_tag_runtime for version and any other information for now.
      "redis_hash_tags_runtime": [
      ],  # Deciding hash tag for every bucket for now, Note that the hash tag must be wrapped in curly braces {}.
      "expire_model_tag_in_seconds": 604800,
      # To eliminate unwanted model versions in Redis to ensure sufficient storage space.
      # It will not take effect if it is less than zero.
      "table_store_mode": 1,
      # Saving and restoring table into ensor in TF savedmodel variable file, table_store_mode = 0;
      # Saving and restoring table into redis rdb file in model_lib_abs_dir, table_store_mode = 1;
      # Saving and restoring nothing, keeping data in redis servers, table_store_mode = 2.
      "model_lib_abs_dir": "/tmp/"
      # if table_store_mode equals 1, then it will try to save or resoter table
      # from model_lib_abs_dir which has been mounted in system
  }

  def __init__(
      self,
      key_dtype,
      value_dtype,
      default_value,
      name="RedisTable",
      checkpoint=False,
      config=None,
      device='',
  ):
    """
      Creates an empty `RedisTable` object.

      Creates a redis table through OS envionment variables,
      the type of its keys and values are specified by key_dtype
      and value_dtype, respectively.

      Args:
        key_dtype: the type of the key tensors.
        value_dtype: the type of the value tensors.
        default_value: The value to use if a key is missing in the table.
        name: A name for the operation (optional, usually it's embedding table name).
        checkpoint: if True, the contents of the table are saved to and restored
          from a Redis binary dump files according to the directory "[model_lib_abs_dir]/[model_tag]/[name].rdb".
          If `shared_name` is empty for a checkpointed table, it is shared using the table node name.

      Returns:
        A `RedisTable` object.

      Raises:
        ValueError: If checkpoint is True and no name was specified.
    """
    self._default_value = ops.convert_to_tensor(default_value,
                                                dtype=value_dtype)
    self._value_shape = self._default_value.get_shape()
    self._checkpoint = checkpoint
    self._key_dtype = key_dtype
    self._value_dtype = value_dtype
    self._device = device
    self._name = name
    self._embedding_name = (self._name.split('_mht_', 1))[0]
    self._config = config

    self.redis_config_file_exist = False
    self.redis_config_file_create = False

    if self._config.redis_config_abs_dir_env:
      if self._config.redis_config_abs_dir_env in os.environ:
        self._config.redis_config_abs_dir = os.getenv(
            self._config.redis_config_abs_dir_env)
      else:
        raise ValueError(
            "Config redis_config_abs_dir_env in RedisTableConfig is not None, but can not find the "
            + self._config.redis_config_abs_dir_env +
            " in system environment variable.")
      self.redis_config_file_exist = os.path.exists(
          self._config.redis_config_abs_dir)
      if self.redis_config_file_exist == False:
        raise ValueError(
            "Config redis_config_abs_dir_env in RedisTableConfig is not None, but the FILE which path stored in environment variable "
            + self._config.redis_config_abs_dir_env + " DOES NOT EXIST.")
    elif self._config.redis_config_abs_dir_env is None and "TFRA_REDIS_CONFIG_PATH" in os.environ:
      self._config.redis_config_abs_dir = os.getenv("TFRA_REDIS_CONFIG_PATH")
      self.redis_config_file_exist = os.path.exists(
          self._config.redis_config_abs_dir)
      warnings.warn(
          "TFRA-Redis try to use environment variable TFRA_REDIS_CONFIG_PATH regardless redis_config_abs_dir in RedisTableConfig."
      )
      if self.redis_config_file_exist == False:
        raise ValueError(
            "environment variable TFRA_REDIS_CONFIG_PATH exists, but the FILE which path stored in TFRA_REDIS_CONFIG_PATH DOES NOT EXIST. Please create a FILE in the corresponding path or delete the environment variable TFRA_REDIS_CONFIG_PATH."
        )
    elif self._config.redis_config_abs_dir_env is None and "TFRA_REDIS_CONFIG_PATH" not in os.environ and self._config.redis_config_abs_dir:
      self.redis_config_file_exist = os.path.exists(
          self._config.redis_config_abs_dir)
      if self.redis_config_file_exist == False:
        raise ValueError(
            "Config redis_config_abs_dir in RedisTableConfig is not None and redis_config_abs_dir_env is None, but the FILE "
            + self._config.redis_config_abs_dir +
            " which path is redis_config_abs_dir DOES NOT EXIST.")
    elif self._config.redis_config_abs_dir_env is None and "TFRA_REDIS_CONFIG_PATH" not in os.environ and self._config.redis_config_abs_dir is None:
      self.redis_config_file_create = True
      self._config.redis_config_abs_dir = "/tmp/tmp_TFRA_Redis_config_file.json"
      warnings.warn(
          "Both redis_config_abs_dir_env and redis_config_abs_dir in RedisTableConfig are None, now creating a temporary config file in /tmp/tmp_TFRA_Redis_config_file.json."
      )
    else:
      raise ValueError(
          "TFRA-Redis didn't get the correct RedisTableConfig class initial parameter."
      )

    if self.redis_config_file_create == True and self.redis_config_file_exist == False:
      with open(self._config.redis_config_abs_dir, 'w+',
                encoding='utf-8') as f0:
        fcntl.flock(f0, fcntl.LOCK_EX)
        f0.write(
            json.dumps(self.default_redis_params, indent=2, ensure_ascii=True))
        fcntl.flock(f0, fcntl.LOCK_UN)
    else:
      with open(self._config.redis_config_abs_dir, 'r', encoding='utf-8') as f0:
        fcntl.flock(f0, fcntl.LOCK_EX)
        params_load = json.load(f0)
        fcntl.flock(f0, fcntl.LOCK_UN)
        self._redis_params = self.default_redis_params.copy()
        for k in self._redis_params.keys():
          if k in params_load:
            self._redis_params[k] = params_load[k]
      with open(self._config.redis_config_abs_dir, 'w', encoding='utf-8') as f1:
        fcntl.flock(f1, fcntl.LOCK_EX)
        f1.write(json.dumps(self._redis_params, indent=2, ensure_ascii=True))
        fcntl.flock(f1, fcntl.LOCK_UN)

    self._shared_name = None
    if context.executing_eagerly():
      # TODO(allenl): This will leak memory due to kernel caching by the
      # shared_name attribute value (but is better than the alternative of
      # sharing everything by default when executing eagerly; hopefully creating
      # tables in a loop is uncommon).
      # TODO(rohanj): Use context.shared_name() instead.
      self._shared_name = "table_%d" % (ops.uid(),)
    super(RedisTable, self).__init__(key_dtype, value_dtype)

    self._resource_handle = self._create_resource()
    if checkpoint:
      _ = RedisTable._Saveable(self, name)
      if not context.executing_eagerly():
        self.saveable = RedisTable._Saveable(
            self,
            name=self._resource_handle.op.name,
            full_name=self._resource_handle.op.name,
        )
        ops.add_to_collection(ops.GraphKeys.SAVEABLE_OBJECTS, self.saveable)
      else:
        self.saveable = RedisTable._Saveable(self, name=name, full_name=name)

  def _create_resource(self):
    # The table must be shared if checkpointing is requested for multi-worker
    # training to work correctly. Use the node name if no shared_name has been
    # explicitly specified.
    use_node_name_sharing = self._checkpoint and self._shared_name is None

    with ops.device(self._device):
      table_ref = redis_table_ops.tfra_redis_table_of_tensors(
          shared_name=self._shared_name,
          use_node_name_sharing=use_node_name_sharing,
          key_dtype=self._key_dtype,
          value_dtype=self._value_dtype,
          value_shape=self._default_value.get_shape(),
          embedding_name=self._embedding_name,
          redis_config_abs_dir=self._config.redis_config_abs_dir,
          redis_config_abs_dir_env=self._config.redis_config_abs_dir_env)

    if context.executing_eagerly():
      self._table_name = None
    else:
      self._table_name = table_ref.op.name.split("/")[-1]
    return table_ref

  @property
  def name(self):
    return self._table_name

  def size(self, name=None):
    """
      Compute the number of elements in this table.

      Args:
        name: A name for the operation (optional).

      Returns:
        A scalar tensor containing the number of elements in this table.
    """
    with ops.name_scope(name, "%s_Size" % self.name, [self.resource_handle]):
      with ops.colocate_with(self.resource_handle):
        return redis_table_ops.tfra_redis_table_size(self.resource_handle)

  def remove(self, keys, name=None):
    """
      Removes `keys` and its associated values from the table.

      If a key is not present in the table, it is silently ignored.

      Args:
        keys: Keys to remove. Can be a tensor of any shape. Must match the table's
          key type.
        name: A name for the operation (optional).

      Returns:
        The created Operation.

      Raises:
        TypeError: when `keys` do not match the table data types.
    """
    if keys.dtype != self._key_dtype:
      raise TypeError("Signature mismatch. Keys must be dtype %s, got %s." %
                      (self._key_dtype, keys.dtype))

    with ops.name_scope(
        name,
        "%s_lookup_table_remove" % self.name,
        (self.resource_handle, keys, self._default_value),
    ):
      op = redis_table_ops.tfra_redis_table_remove(self.resource_handle, keys)

    return op

  def clear(self, name=None):
    """
      clear all keys and values in the table.

      Args:
        name: A name for the operation (optional).

      Returns:
        The created Operation.
    """
    with ops.name_scope(name, "%s_lookup_table_clear" % self.name,
                        (self.resource_handle, self._default_value)):
      op = redis_table_ops.tfra_redis_table_clear(self.resource_handle,
                                                  key_dtype=self._key_dtype,
                                                  value_dtype=self._value_dtype)

    return op

  def lookup(self,
             keys,
             dynamic_default_values=None,
             return_exists=False,
             name=None):
    """
      Looks up `keys` in a table, outputs the corresponding values.

      The `default_value` is used for keys not present in the table.

      Args:
        keys: Keys to look up. Can be a tensor of any shape. Must match the
          table's key_dtype.
        dynamic_default_values: The values to use if a key is missing in the
          table. If None (by default), the static default_value
          `self._default_value` will be used.
        return_exists: if True, will return a additional Tensor which indicates
          if or not keys are existing in the table.
        name: A name for the operation (optional).

      Returns:
        A tensor containing the values in the same shape as `keys` using the
          table's value type.
        exists:
          A bool type Tensor of the same shape as `keys` which indicates
            if keys are existing in the table.
            Only provided if `return_exists` is True.

      Raises:
        TypeError: when `keys` do not match the table data types.
    """
    with ops.name_scope(
        name,
        "%s_lookup_table_find" % self.name,
        (self.resource_handle, keys, self._default_value),
    ):
      keys = ops.convert_to_tensor(keys, dtype=self._key_dtype, name="keys")
      with ops.colocate_with(self.resource_handle):
        if return_exists:
          values, exists = redis_table_ops.tfra_redis_table_find_with_exists(
              self.resource_handle,
              keys,
              dynamic_default_values
              if dynamic_default_values is not None else self._default_value,
          )
        else:
          values = redis_table_ops.tfra_redis_table_find(
              self.resource_handle,
              keys,
              dynamic_default_values
              if dynamic_default_values is not None else self._default_value,
          )
    return (values, exists) if return_exists else values

  def insert(self, keys, values, name=None):
    """
      Associates `keys` with `values`.

      Args:
        keys: Keys to insert. Can be a tensor of any shape. Must match the table's
          key type.
        values: Values to be associated with keys. Must be a tensor of the same
          shape as `keys` and match the table's value type.
        name: A name for the operation (optional).

      Returns:
        The created Operation.

      Raises:
        TypeError: when `keys` or `values` doesn't match the table data
          types.
    """
    with ops.name_scope(
        name,
        "%s_lookup_table_insert" % self.name,
        [self.resource_handle, keys, values],
    ):
      keys = ops.convert_to_tensor(keys, self._key_dtype, name="keys")
      values = ops.convert_to_tensor(values, self._value_dtype, name="values")
      with ops.colocate_with(self.resource_handle):
        # pylint: disable=protected-access
        op = redis_table_ops.tfra_redis_table_insert(self.resource_handle, keys,
                                                     values)
    return op

  def accum(self, keys, values_or_deltas, exists, name=None):
    """Associates `keys` with `values`.

      Args:
        keys: Keys to accmulate. Can be a tensor of any shape.
          Must match the table's key type.
        values_or_deltas: values to be associated with keys. Must be a tensor of
          the same shape as `keys` and match the table's value type.
        exists: A bool type tensor indicates if keys already exist or not.
          Must be a tensor of the same shape as `keys`.
        name: A name for the operation (optional).

      Returns:
        The created Operation.

      Raises:
        TypeError: when `keys` or `values` doesn't match the table data
          types.
    """
    with ops.name_scope(
        name,
        "%s_lookup_table_accum" % self.name,
        [self.resource_handle, keys, values_or_deltas],
    ):
      keys = ops.convert_to_tensor(keys, self._key_dtype, name="keys")
      values_or_deltas = ops.convert_to_tensor(values_or_deltas,
                                               self._value_dtype,
                                               name="values_or_deltas")
      exists = ops.convert_to_tensor(exists, dtypes.bool, name="exists")
      with ops.colocate_with(self.resource_handle):
        # pylint: disable=protected-access
        op = redis_table_ops.tfra_redis_table_accum(self.resource_handle, keys,
                                                    values_or_deltas, exists)
    return op

  def export(self, name=None):
    """
      Returns nothing in Redis Implement. It will dump some binary files
      to model_lib_abs_dir.

      Args:
        name: A name for the operation (optional).

      Returns:
        A pair of tensors with the first tensor containing all keys and the
          second tensors containing all values in the table.
    """
    with ops.name_scope(name, "%s_lookup_table_export_values" % self.name,
                        [self.resource_handle]):
      with ops.colocate_with(self.resource_handle):
        (
            exported_keys,
            exported_values,
        ) = redis_table_ops.tfra_redis_table_export(self.resource_handle,
                                                    self._key_dtype,
                                                    self._value_dtype)
    return exported_keys, exported_values

  def save_to_file_system(self,
                          dirpath,
                          file_name=None,
                          dirpath_env='TFRA_SAVED_KV',
                          append_to_file=False,
                          buffer_size=4194304,
                          name=None):
    """
    Returns an operation to save the keys and values in table to dirpath. 
    The keys and values will be stored in FileSystem, rewrited or appended to the filepath.
    Args:
      dirpath: A directory path to save the table.
      dirpath_env: A environment variable stored a path to save the table, which priority higher than dirpath.
      file_name: User custom file name for key/value prefix file name, default is self._name.
      buffer_size: Number of keys in write buffer to file.
      append_to_file: If true, operation will append data to the file but not write a new one.
      name: Name for the operation.
    Returns:
      An operation to save the table.
    """
    with ops.name_scope(name, "%s_save_table" % self.name,
                        [self.resource_handle]):
      with ops.colocate_with(None, ignore_existing=True):
        return redis_table_ops.tfra_redis_table_save_to_file_system(
            self.resource_handle,
            dirpath=dirpath,
            file_name=file_name if file_name else self._name,
            key_dtype=self._key_dtype,
            value_dtype=self._value_dtype,
            dirpath_env=dirpath_env,
            append_to_file=append_to_file,
            buffer_size=buffer_size)

  def load_from_file_system(self,
                            dirpath,
                            file_name=None,
                            dirpath_env='TFRA_SAVED_KV',
                            load_entire_dir=False,
                            buffer_size=4194304,
                            name=None):
    """
    Returns an operation to load keys and values to table from
    FileSystem. The keys and values files are generated from `save_to_file_system`.
    Args:
      dirpath: A directory path stored the table keys and values.
      dirpath_env: A environment variable stored a path to load the table, which priority higher than dirpath.
      file_name: User custom file name for key/value prefix file name, default is self._name.
      buffer_size: Number of keys in read buffer from file.
      load_entire_dir: If true, operation will load all key value files in the dirpath regardless partition.
      name: Name for the operation.
    Returns:
      An operation to load keys and values to table from FileSystem.
    """
    with ops.name_scope(name, "%s_load_table" % self.name,
                        [self.resource_handle]):
      with ops.colocate_with(None, ignore_existing=True):
        return redis_table_ops.tfra_redis_table_load_from_file_system(
            self.resource_handle,
            dirpath=dirpath,
            file_name=file_name if file_name else self._name,
            key_dtype=self._key_dtype,
            value_dtype=self._value_dtype,
            dirpath_env=dirpath_env,
            load_entire_dir=load_entire_dir,
            buffer_size=buffer_size)

  def _gather_saveables_for_checkpoint(self):
    """For object-based checkpointing."""
    # full_name helps to figure out the name-based Saver's name for this saveable.
    full_name = self._table_name
    return {
        "table":
            functools.partial(
                RedisTable._Saveable,
                table=self,
                name=self._name,
                full_name=full_name,
            )
    }

  class _Saveable(BaseSaverBuilder.SaveableObject):
    """SaveableObject implementation for RedisTable."""

    def __init__(self, table, name, full_name=""):
      tensors = table.export()
      specs = [
          BaseSaverBuilder.SaveSpec(tensors[0], "", name + "-keys"),
          BaseSaverBuilder.SaveSpec(tensors[1], "", name + "-values"),
      ]
      # pylint: disable=protected-access
      super(RedisTable._Saveable, self).__init__(table, specs, name)
      self._restore_name = table._name

    def restore(self, restored_tensors, restored_shapes, name=None):
      del restored_shapes  # unused
      # pylint: disable=protected-access
      with ops.name_scope(name, "%s_table_restore" % self._restore_name):
        with ops.colocate_with(self.op.resource_handle):
          return redis_table_ops.tfra_redis_table_import(
              self.op.resource_handle,
              restored_tensors[0],
              restored_tensors[1],
          )


ops.NotDifferentiable(prefix_op_name("RedisTableOfTensors"))