Merge pull request #42 from pockerman/add_sarsa_semi_gradient

Refactorings

Author: pockerman

src/algorithms/q_learning.py

@@ -89,7 +89,7 @@ def play(self, env: Env, stop_criterion: Criterion) -> None:
             env.step(action=action)
             total_dist = env.total_current_distortion()
 
-    def train(self, env: Env, **options) -> tuple:
+    def on_episode(self, env: Env, **options) -> tuple:
 
         # episode score
         episode_score = 0

src/algorithms/trainer.py

@@ -74,7 +74,7 @@ def train(self):
             ignore = self.env.reset()
 
             # train for a number of iterations
-            episode_score, total_distortion, n_itrs = self.agent.train(self.env)
+            episode_score, total_distortion, n_itrs = self.agent.on_episode(self.env)
 
             print("{0} Episode score={1}, episode total distortion {2}".format(INFO, episode_score, total_distortion / n_itrs))
 

src/spaces/column_type.py

@@ -0,0 +1,17 @@
+"""
+Simple enumeration of column types.
+This is similar to the ARX software.
+See the ARX documentation at:
+https://arx.deidentifier.org/wp-content/uploads/javadoc/current/api/org/deidentifier/arx/AttributeType.html
+"""
+
+import enum
+
+
+class ColumnType(enum.IntEnum):
+
+    INVALID_TYPE = 0
+    IDENTIFYING_ATTRIBUTE = 1
+    SENSITIVE_ATTRIBUTE = 2
+    INSENSITIVE_ATTRIBUTE = 3
+    QUASI_IDENTIFYING_ATTRIBUTE = 4

src/spaces/discrete_state_environment.py

@@ -4,69 +4,19 @@
 """
 
 import copy
-import enum
 import numpy as np
 from pathlib import Path
-import pandas as pd
-import torch
-from typing import NamedTuple, Generic, Optional, TypeVar, List
+from typing import TypeVar, List
 import multiprocessing as mp
 
 from src.spaces.actions import ActionBase, ActionType
-from src.utils.string_distance_calculator import StringDistanceType, TextDistanceCalculator
-from src.utils.numeric_distance_type import NumericDistanceType
-from src.utils.numeric_distance_calculator import NumericDistanceCalculator
+from src.spaces.time_step import TimeStep, StepType
 
 DataSet = TypeVar("DataSet")
 RewardManager = TypeVar("RewardManager")
 ActionSpace = TypeVar("ActionSpace")
 DistortionCalculator = TypeVar('DistortionCalculator')
 
-_Reward = TypeVar('_Reward')
-_Discount = TypeVar('_Discount')
-_Observation = TypeVar('_Observation')
-
-
-class StepType(enum.IntEnum):
-    """
-      Defines the status of a `TimeStep` within a sequence.
-      """
-
-    # Denotes the first `TimeStep` in a sequence.
-    FIRST = 0
-
-    # Denotes any `TimeStep` in a sequence that is not FIRST or LAST.
-    MID = 1
-
-    # Denotes the last `TimeStep` in a sequence.
-    LAST = 2
-
-    def first(self) -> bool:
-        return self is StepType.FIRST
-
-    def mid(self) -> bool:
-        return self is StepType.MID
-
-    def last(self) -> bool:
-        return self is StepType.LAST
-
-
-class TimeStep(NamedTuple, Generic[_Reward, _Discount, _Observation]):
-    step_type: StepType
-    info: dict
-    reward: Optional[_Reward]
-    discount: Optional[_Discount]
-    observation: _Observation
-
-    def first(self) -> bool:
-        return self.step_type == StepType.FIRST
-
-    def mid(self) -> bool:
-        return self.step_type == StepType.MID
-
-    def last(self) -> bool:
-        return self.step_type == StepType.LAST
-
 
 class DiscreteEnvConfig(object):
     """
@@ -79,8 +29,6 @@ def __init__(self) -> None:
         self.reward_manager: RewardManager = None
         self.average_distortion_constraint: float = 0.0
         self.gamma: float = 0.99
-        # self.string_column_distortion_type: StringDistanceType = StringDistanceType.INVALID
-        # self.numeric_column_distortion_metric_type: NumericDistanceType = NumericDistanceType.INVALID
         self.n_states: int = 10
         self.min_distortion: float = 0.4
         self.max_distortion: float = 0.7
@@ -115,6 +63,10 @@ def __init__(self, env_config: DiscreteEnvConfig) -> None:
         self.column_visits = {}
         self.create_bins()
 
+    @property
+    def columns_attribute_types(self) -> dict:
+        return self.config.data_set.columns_attribute_types
+
     @property
     def action_space(self):
         return self.config.action_space

src/spaces/time_step.py

@@ -0,0 +1,55 @@
+"""
+
+"""
+
+import enum
+from typing import NamedTuple, Generic, Optional, TypeVar
+
+_Reward = TypeVar('_Reward')
+_Discount = TypeVar('_Discount')
+_Observation = TypeVar('_Observation')
+
+
+class StepType(enum.IntEnum):
+    """
+      Defines the status of a `TimeStep` within a sequence.
+      """
+
+    # Denotes the first `TimeStep` in a sequence.
+    FIRST = 0
+
+    # Denotes any `TimeStep` in a sequence that is not FIRST or LAST.
+    MID = 1
+
+    # Denotes the last `TimeStep` in a sequence.
+    LAST = 2
+
+    def first(self) -> bool:
+        return self is StepType.FIRST
+
+    def mid(self) -> bool:
+        return self is StepType.MID
+
+    def last(self) -> bool:
+        return self is StepType.LAST
+
+
+class TimeStep(NamedTuple, Generic[_Reward, _Discount, _Observation]):
+    step_type: StepType
+    info: dict
+    reward: Optional[_Reward]
+    discount: Optional[_Discount]
+    observation: _Observation
+
+    def first(self) -> bool:
+        return self.step_type == StepType.FIRST
+
+    def mid(self) -> bool:
+        return self.step_type == StepType.MID
+
+    def last(self) -> bool:
+        return self.step_type == StepType.LAST
+
+    @property
+    def done(self) -> bool:
+        return self.last()