org.deeplearning4j.gym

Class Client<O,A,AS extends ActionSpace<A>>

java.lang.Object

org.deeplearning4j.gym.Client<O,A,AS>

Type Parameters:

O - Observation type

A - Action type

AS - Action Space type

public class Client<O,A,AS extends ActionSpace<A>>
extends Object

Author:

rubenfiszel (ruben.fiszel@epfl.ch) on 7/6/16. A client represent an active connection to a specific instance of an environment on a rl4j-http-api server. for API specification

See Also:

https://github.com/openai/gym-http-api#api-specification

Field Summary

Fields

Modifier and Type	Field and Description
static String	ACTION_SPACE
static String	CLOSE
static String	ENVS_ROOT
static String	MONITOR_CLOSE
static String	MONITOR_START
static String	OBSERVATION_SPACE
static String	RESET
static String	SHUTDOWN
static String	STEP
static String	UPLOAD
static String	V1_ROOT

Constructor Summary

Constructors

Constructor and Description
Client()

Method Summary

Modifier and Type	Method and Description
ActionSpace	getActionSpace()
Set<String>	listAll()
static Set<String>	listAll(String url)
void	monitorClose() Flush all monitor data to disk
void	monitorStart(String directory, boolean force, boolean resume) Start monitoring.
O	reset() Reset the state of the environment and return an initial observation.
void	serverShutdown() Shutdown the server at the same url than this client
static void	serverShutdown(String url) Shutdown the server at the url
StepReply<O>	step(A action) Step the environment by one action
void	upload(String trainingDir, String apiKey) Upload monitoring data to OpenAI servers.
void	upload(String trainingDir, String apiKey, String algorithmId) Upload monitoring data to OpenAI servers.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

V1_ROOT

public static String V1_ROOT

ENVS_ROOT

public static String ENVS_ROOT

MONITOR_START

public static String MONITOR_START

MONITOR_CLOSE

public static String MONITOR_CLOSE

CLOSE

public static String CLOSE

RESET

public static String RESET

SHUTDOWN

public static String SHUTDOWN

UPLOAD

public static String UPLOAD

STEP

public static String STEP

OBSERVATION_SPACE

public static String OBSERVATION_SPACE

ACTION_SPACE

public static String ACTION_SPACE

Constructor Detail

Client

public Client()

Method Detail

listAll

public static Set<String> listAll(String url)

Parameters:

url - url of the server

Returns:

set of all environments running on the server at the url

serverShutdown

public static void serverShutdown(String url)

Shutdown the server at the url

Parameters:

url - url of the server

listAll

public Set<String> listAll()

Returns:

set of all environments running on the same server than this client

step

public StepReply<O> step(A action)

Step the environment by one action

Parameters:

action - action to step the environment with

Returns:

the StepReply containing the next observation, the reward, if it is a terminal state and optional information.

reset

public O reset()

Reset the state of the environment and return an initial observation.

Returns:

initial observation

monitorStart

public void monitorStart(String directory,
                         boolean force,
                         boolean resume)

Start monitoring.

Parameters:

directory - path to directory in which store the monitoring file

force - clear out existing training data from this directory (by deleting every file prefixed with "openaigym.")

resume - retain the training data already in this directory, which will be merged with our new data

monitorClose

public void monitorClose()

Flush all monitor data to disk

upload

public void upload(String trainingDir,
                   String apiKey,
                   String algorithmId)

Upload monitoring data to OpenAI servers.

Parameters:

trainingDir - directory that contains the monitoring data

apiKey - personal OpenAI API key

algorithmId - an arbitrary string indicating the paricular version of the algorithm (including choices of parameters) you are running.

upload

public void upload(String trainingDir,
                   String apiKey)

Upload monitoring data to OpenAI servers.

Parameters:

trainingDir - directory that contains the monitoring data

apiKey - personal OpenAI API key

serverShutdown

public void serverShutdown()

Shutdown the server at the same url than this client

getActionSpace

public ActionSpace getActionSpace()